Parallelism and Orders of Signification (Parallelism Dynamics I)

Oral Tradition, 31/2 (2017): 425-484 Parallelism and Orders of Signification (Parallelism Dynamics I) Frog Discussions 1 of parallelism in verbal art have customarily focused on semantic and grammatical parallelism between adjacent verses or equivalent units, “similarities between discourse segments that are sequentially juxtaposed” (Urban 1991:60).2 It is easy to get the impression that parallelism mainly concerns dyadic pairs that form couplets, even if it is common to acknowledge extended series of parallel verses,3 figures like cross parallelism or chiasmus,4 and parallelism between verses separated by one or a few lines of intermediate text.5 However, Gerard Manley Hopkins (1959 [1865]:84) reflected that “perhaps we shall be right to 1 Research presented here was completed within the framework of the project “The Song of Lemminkäinen: A Finno-Karelian Epic in Parallax Perspective,” funded by the Kalevala Society. It has been developed out of the first half of the paper “Parallelism, Mode, Medium and Orders of Representation,” presented at the seminarworkshop Parallelism in Verbal Art and Performance, held 26th-27th May 2014 in Helsinki, Finland. The approach presented here is largely an outcome of discussions with other scholars leading up to, during, and following that event, among whom I would particularly like to thank Richard Bauman, James J. Fox, Karina Lukin, and Eila Stepanova. I also appreciate the valuable comments and suggestions of my peer-reviewers and the stimulating discussions with contributors to the present collection. I would especially like to thank John Zemke, for his generous feedback, tireless editorial work, and boundless patience that have greatly strengthened this article. 2 See, for example, James J. Fox’s classic article “Roman Jakobson and the Comparative Study of Parallelism” (1977), which opens by introducing the potential breadth of the concept (59-60) and then discusses parallelism almost exclusively in terms of adjacent verses (60-81; a revised version of this article appears as Fox 2014:19-40). In his section on “Criteria for defining ‘canonical parallelism,’” Fox (1977:77-80 and 2014:37-40) emphasizes the operation of word pairs in parallelism without linking to his preceding discussion of extended parallel series and complex structures (1977:73-77 and 2014:32-37). He then concludes by referring to “the oral poet, speaking in pairs” (1977:80 and 2014:40), a phrase that became almost emblematic of parallelism through use in the title of the influential collection To Speak in Pairs (Fox 1988). 3 For example, Barbara Johnstone (1991) distinguishes “[c]anonical parallelism and couplet structure” (21-27) of the A1 A2 B1 B2 type from “[t]he rhetoric of the series” (27-32) such as an A1 A2 A3 . . . An type, implying that a series of more than two members should not be conflated with “canonical parallelism.” Johnstone’s distinction may seem exaggerated, but it represents a way of reading Fox’s “Criteria for defining ‘canonical parallelism.’” 4 Cross parallelism describes an A B A B pattern while chiasmus describes an A B B A pattern (see, 1 1 2 2 1 1 2 2 for example, Berlin 1985:83-88; Holm, this volume). 5 This sort of proximate parallelism forms patterns like A X A X or A B C X X X A B C , although 1 2 1 1 1 2 2 2 these often shift into the background of discussion (see, for example, Austerlitz 1958:47-48; Fabb 2015:160-61 on Lukin 2014:122-29). 426 FROG say all artifice, reduces to the principle of parallelism,” a perspective that is today more familiar from Roman Jakobson’s view that the “recurrent returns” of parallelism are “the essence of poetic artifice” “at every level of language” (1981 [1966]:98). From this perspective, “recurrent returns” at the level of sounds, words, syntax, morphology, semantic units, and so forth, all fall under the aegis of “parallelism;”6 alliteration and rhyme are “only a particular, condensed case of a much more general” phenomenon of parallelism (1981 [1960]:39). It may seem peculiar to some readers that parallel verses that reiterate the same information should be compared to rhyme, but this line of thinking was current at least as early as the seventeenth century,7 and is built into the German term Gedankenreim (“thought-rhyme”) for semantic parallelism. In this article I am interested in theorizing and formalizing this broader view of parallelism with a focus on verbal art, and in developing methodological tools for its analysis. Research on verbal art tends to look at how a tradition works in terms of the surface levels of communication—what is heard and seen in a performance situation. Everything involved in the verbal part of communication tends to be viewed more or less exclusively at the level of language. A “theme” in the sense of Oral-Formulaic Theory is thus recognized as a unit of narration but is treated as a unit of language, a unit of utterance, rather than as something mediated by language (for example, Foley 1999:83-86). Here, a central concern is to peel apart the layers involved in communication, separating sounds from the words they mediate, words from the images and motifs they mediate, and these from the signs that they mediate, and so on. When these layers are distinguished, parallelism can be seen as “recurrent returns” of elements within the same layer in relation to the metered frames of units of utterance. Parallelism in each layer can then be looked at in relation to what is happening in other layers, how and whether parallelism in language relates to motifs or themes, and so on. The approach to parallelism presented here has developed from my work on identifying and understanding units that are meaningful in a tradition, whether they are narrated through language, allusively referred to in conversation, represented iconographically, or enacted. There is nothing profound in observing that a motif is not, strictly speaking, a unit of language but rather something that can be communicated using language or in some other way. Working across language, material culture, and embodied activity motivated me to differentiate types of units in order to address how they interact in use. It became necessary to examine how these units are affected or shaped by a medium such as language and may become interfaced with it. Here, the focus is on verbal art and meaningful units mediated through language in verbal art. However, the platform for approaching the units and their relations has a much broader background. Thus, the definitions of “motif” and other units of tradition described here have 6 For studies addressing parallelism from this broader perspective, see for example Berlin (1985); Johnstone (1991); see also contributions by Stepanova, and Turpin in this volume. 7 A century before Robert Lowth (1753:180) famously introduced the term parallelismus membrorum (“parallelism of members”), Joseph Mede (1653:114) observed that Hæbræa poesis rhythmum habuit, non in sono, nisi fortuito, sed in sensu; idem vel simile, diversa phrase reduplicans (“The rhythm of Hebrew poetry was not in sound, except by chance, but in sense, reduplicating the same or similar phrases”). Jean des Champs refers to parallelism as an espèce de rime (“species of rhyme”), a rime du sens (“rhyme of sense”) (1754:269). In Latin discussions, rhythmus sensus and rhythmus soni became paired terms accompanied by the French (Ullholm/ Aurivillius 1758:8; in the discussion of Kalevala-meter poetry in Porthan’s 1766:22), when rime and rhythmus referred to poetic forms that lacked meter in a narrow, classical sense. PARALLELISM DYNAMICS I 427 been developed to analyse symbols and their arrangements in Iron-Age burials, how the same units can be found across medieval poetry, prose, and iconography, how verbal art may relate to actions, objects, and the environment in a ritual performance, or the anticipated outcome of violating a taboo (see, for example, Frog 2014a:360-79 and 2015a:35-47). I present an overview of what becomes observable in verbal art when language is distinguished from other signs as a medium for their use and communication. The discussion is organized with an introduction to terms and concepts in §1. A central tool introduced is John W. Du Bois’ “diagraph” analysis (2007:159-62 and 2014:362-63, 376-78). This tool was developed for studying emergent parallelism in co-produced conversation but is applied here for the analysis of verse parallelism and also adapted from use with linguistic units to semantic and symbolic units of expression. The discussion is illustrated through oral kalevalaic epic poetry, although the theory and methodology are more generally applicable. Kalevalaic epic poetry has certain formal qualities and conventions that facilitate discussion, while illustration through a single poetic system also makes discussion more coherent. This tradition is introduced in §2, with relevant consideration of aspects of poetic form, language, units of composition, and some germane remarks on verse parallelism, as well as outlines of the plots of the two epics from which most examples are drawn. In §3 and §4, parallelism based on images and motifs as minimal units of tradition is discussed, followed by what is referred to as higher order parallelism in §5, and a brief conclusion in §6. 1. Some Terms, Tools and Concepts 1.1. Parallelism, Resonance, and Reproduction Basically, parallelism is a type of repetition with difference in which parallel units are perceived as parallel members of groups (Cureton 1992:263). More specifically, parallelism is here defined as a perceivable quality of recurrence of sameness or similarity in commensurable units that co-occurs with difference in relation to a metered frame. The term “metered frame” is taken from Michael Silverstein (1984:183). In an approach to discourse consistent with ethnopoetic analysis,8 Silverstein proposes that a unit of utterance presents a “metered frame” in relation to which subsequent co-occurring utterances are perceived (see also Johnstone 1991:33). In Silverstein’s discussion, the formal correlation of metered frames and how units of language relate to them is not addressed further. As a consequence, Silverstein’s parallelism converges 8 For an example of what I refer to as ethnopoetic analysis, see Hymes (1981:200-59); see also Tedlock (1983:285-338) and Blommaert (2006:232-41). Ethnopoetics is often seen as a major step forward in Western thought, but during the Classical period, roughly two thousand years before Dell Hymes and Dennis Tedlock were writing, rhythm was already recognized as fundamental to a broader range of linguistic behavior, whereas poetry was more narrowly qualified by formalized meter. As Quintilian points out: contra nihil quod est prorsa scriptum non redigi possit in quaedam versiculorum genera uel in membra (“certainly there is nothing written in prose that cannot be reduced to some sort of verses or indeed parts of verses”) (Institutio Oratoria IX.iv.lii). 428 FROG with deixis: it is inclusive of anything that refers back to something previously said.9 Here, commensurability of units in relation to a metered frame establishes a formal condition to parallelism (see also Du Bois 2014:370). Silverstein’s concept of “metered frame” remains a significant contribution to the discussion of parallelism and is further developed here. Emergent metered frames can be viewed in the light of Reuven Tsur’s emphasis on the etymological background of “articulation” as “jointed, separated into well-shaped pieces” (1992:150). Entextualization construes utterance as units: units are “articulated,” distinguishing them from one another as units. Articulation may be through a formalized meter with regularly recurrent rhythms such as Kalevala-meter introduced below. It may be through units like the “strings” of Karelian laments that are marked by alliteration and melodic contours but are extremely variable in length (Stepanova, this volume). Articulation may also be ethnopoetically marked by breaths, expletives, syntax, melodic phrases, and so forth in aesthetically unmarked discourse as in a stand-up comedy routine (Lindfors, this volume). Each such unit presents a metered frame that allows it to be correlated with other co-occurring units. Du Bois (2014:397-400) argues that parallelism is a phenomenon of syntax, following Charles W. Morris’ (1971 [1938]:22) definition of “the syntactical dimension of semiosis” as “the formal relation of signs to one another.” The syntactic role of parallelism is transparent in cases of ellipsis. In example (12) below, for instance, there is no verb in the second verse of a couplet, but it is inferred through the parallelism with the preceding verse: “veri vuodi vembelestä / rasva rahkehen nenästä” (SKVR I1 163a.6-7; following manuscript orthography) (“Blood ran from the shaft-bow / Fat from the trace’s end”). Developing a sort of Grammatik des Parallelismus (“grammar of parallelism”) was already a concern of Wolfgang Steinitz (1934:xii) in his seminal study of parallelism in kalevalaic poetry. Here, the view of parallelism as a syntactic phenomenon is extended to relations between non-linguistic signs as well. In order to address “parallelism” as a term and concept, Du Bois has formalized the complementary terms “resonance” and “reproduction.” Du Bois (2014:372) defines “resonance” as: the catalytic activation of affinities across [units of utterance]. 10 Resonance is a property of relations between elements of discourse; as such it cannot be attributed to any element in isolation. It represents a developing process of activation and elaboration of certain aspects of the perceived relationship between comparable linguistic elements. 9 Silverstein’s primary concern in this discussion is how conversational participants organize relationships to one another through utterances that relate to prior utterances. He gives particular attention to deictic terms, lexical repetitions, and equivalency and contrast in semantic content. In his later address of the same material (2004), a metered or metricalized frame is referred to as “a conceptual metrics” with emphasis on what he calls “deictic metricalization” (2004:629). Rather than parallelism, the metricality being brought into focus describes the rhythms of reference in dialogue (in literature, see Hasan 2007:23-32). 10 Du Bois has simply “utterances,” which is consistent with his emphasis on parallelism across utterances by different interlocutors in dialogic engagement. In this context, he seems to correlate utterance with an ethnopoetic line (see Du Bois 2007:371 #5 and 7, 373-74 #8-9). This may simply be a difference in terminology: when addressing oral poetry, treating each metrical line as a discrete utterance seems to atomize the poetry and may be inconsistent with its syntax. PARALLELISM DYNAMICS I 429 Resonance creates relations between utterances while parallelism is a product of generating such relations; parallelism can thus be viewed as a form of resonance (2014:374). The utility of “resonance” as a term and concept is that it provides a broader frame for considering the relationality of expressions among which parallelism can occur. I use “reproduction” to refer to the production of a previously produced signifier in a context where a relationship between the two or more uses is salient (2014:378).11 Reproduction in this sense equates to what is commonly called “repetition” (which may ultimately prove the preferred term in scholarship).12 However, when parallelism itself is viewed as a form of repetition with variation, it is useful to have a complementary term for distinguishing signifiers that recur in parallel members. 1.2. Orders of Signification Primary mode Verbalization Symbolic articulation Higher order 1 Higher order 2 Higher order 3 / • / • / • / • / • / • / • / • / • / • / • / • / • / • / • / • / • / • / • / • /•/•/•/• /•/•/•/• words words words image or motif words image or motif words words words image or motif theme /•/•/•/• words image or motif them me narrrative patteern / episodde ← EP PIC → Fig. 1. Visual illustration of expressions manifesting metered frames at multiple levels in signification simultaneously. The different levels are introduced below. The diagram illustrates that the organizing principles of the primary mode of expression structure the units of language; language is in turn a mode for symbolic articulation, and so forth. Each mode formally conditions the metered frames of the following order of signification in relation to its syntax. Parallelism is a phenomenon of signification. My approach is built on distinguishing different levels in signification. These levels include sounds or script that mediate language, language that mediates other types of signs, signs mediated by those signs, and so on, as in the somewhat idealized illustration of Fig. 1. According to Saussurean semiotics, a sign is made up of a signifier and that which it signifies (Saussure 1967 [1916]:97-103). Words are thus signs: a word is made up of a recognizable signifier and its associated meaning (signified). In Saussure’s terms, the sound through which words are communicated is a signal rather than a sign itself. Words may also mediate signs such as images and motifs. Those images and motifs may, in turn, mediate discrete meaning-bearing units that are also signs. Each level in signification becomes a potential site where parallelism may occur. In this section, the theoretical side of this topic is discussed, along with some of the issues connected with it. Some readers may be satisfied 11 Du Bois includes the recurrence of structures under reproduction as well, which here is addressed as a form of parallelism when the structure recurs with different words. 12 Use of the term “repetition” has been criticized in the discussion of oral traditions, deconstructing it to reveal its undesirable connotations (for example, Foley 1991:56-59). However, the issue is quite messy because the term “repetition” is rarely formalized and it has a wider range of uses than parallelism. Repeated deconstruction of the term leads its connotations to change, and “repetition” will quite possibly be unburdened of baggage before an alternative is widely established. 430 FROG drawing the basic principles from Fig. 1 and moving ahead or first going through the empirical examples and then returning to issues of theory. In terms of processes of semiosis, however, Fig. 1 is highly idealized in higher order signification in particular, requiring some discussion. Two closely related concepts used here are “medium” and “mode” of expression, although concern is almost exclusively with mode. The difference between them is not especially important in the discussion below, but it will be briefly clarified here to avoid confusion. A “medium” is that which mediates; it is a vehicle for signs. Thus, voice is a medium for words, and words, in their turn, can be a medium for other signs. Whereas “medium” here designates a broad category of sign vehicle, “mode”13 denotes a conventional way of using a medium. Thus, dictation and sung performance rely on voice as a common medium as a vehicle for language but they employ different modes. Similarly, language provides a common medium for narration in oral epic verse and an epic’s summary in prose, but these may be very different modes of language use. In order to avoid ambiguity, “primary mode” will refer to the conventional mode (or modes) for mediating language in verbal art. Meter or principles of metricalization are here considered to operate at the level of mode (Frog 2012b:52-54). However, concern here with mode (and medium) and its relationship to parallelism remains simply as a layer of signification within the performance tradition of a variety of verbal art. Shifts between modes and uses of different modes on a situational basis will not be discussed. In an oral form of verbal art, the principles and conventions that structure the primary mode act as determinants on the entextualization of language. For example, phonic requirements of alliteration or syllabic rhythms affect the choice of words and their organization. Because the primary mode is a level of signals as a vehicle for linguistic signs, parallelism manifests as recurrent returns in sounds and their rhythms. These recurrent returns simultaneously form metered frames for units of utterance and occur in relation to them.14 In many oral poetries, phonic parallelism is particularly salient because it is regular and predicted, whether because it is metrically required or formally prominent. Regular meter and/or the reproduction of melodic or intonational structures also manifest as recurrent returns that can form parallelisms. The organized patterning of recurrent rhythms, intonational structures, and phonic patterns 15 in relation to metered frames make the units of utterance more salient. In the continuous flow of performance, parallelism in the primary mode is both part of the articulation of units of utterance and also organizes them as parallel members of groups, producing cohesion so that they are 13 The term “mode” is adapted from Michael Halliday (1978:64, espec.) in Systemic-Functional Linguistics (see also Shore 2015:63-66, 68). Halliday’s use of this term sometimes blurs with medium, for example using it to distinguish oral and written “modes” of communication. 14 Although patterns of rhyme and alliteration are often described as recurrence of the same phoneme, the recurrent sounds are customarily used in the signifiers of different words. In some traditions, these patterns may also occur through figura etymologica, a rhetorical figure that uses different but genetically related words in close enough proximity to be noticeable (see also Stepanova, this volume; compare Johnstone 1991:53-55, 62-71 on “root repetition”). A similar effect may be produced by the reproduction of the same word in different inflectional forms, as in example (12) below. 15 Predictable phonic patterns like alliteration and rhyme have this effect whether they are organized within each unit or recur at the onsets or endings of proximate units. PARALLELISM DYNAMICS I 431 perceived as being more closely related to one another than to preceding and subsequent discourse. Of crucial concern here are the differentiation of language from signs mediated by language and the differentiation of orders of signs. The first concern is to avoid collapsing and conflating language with images and motifs as symbols mediated by language. The second is to provide a platform for understanding the workings of how different symbols interact in meaning production, and to avoid reducing them to formal units in structural hierarchies. In connection with these distinctions, another pair of closely related concepts used here is “level in signification” and “order of signification.” As visually represented in Fig. 1, signification, the production of signs, has several layers or levels. “Level in signification” designates any layer in that process, whether it is comprised of signals, like voice as a medium for language, or signs, such as words of language. “Order of signification” (not to be confused with “order of indexicality”16) refers to a layer of signs in relation to its position in a series of signs mediating other signs. As a rule of thumb, any traditional unit that can be recognized and referred to as meaningful is a sign. Signs that are mediated through language will here be referred to generally as symbols. Drawing on the terminology of John Miles Foley, these symbolic units are described as “integers,” a term that emphasizes that they operate as unitary signifiers with propositional, indexical, or ambiguous significance.17 The first order of signification is the sign system that you access directly through signals; the second is accessed through the first, the third through the second, and so forth. The relations of these orders can be illustrated as: Signals → O1 → O2 → O3 → etc. Language can be mediated by voice or writing. When linguistic signs are produced vocally, voice produces signals from which sequences of sounds like /hors/ are perceived as signifiers of linguistic signs like “horse.” In this way, language is apprehended as the first order of signification, and a mythic image of, say, a winged horse communicated through language would be at the second order (O2). Addressing orders in terms of “first,” “second,” and “third” leads to a stumbling block when written archival sources are introduced because, technically, an orthographic sign system is interposed at the beginning of the series in order to mediate language in the absence of voice. The system of writing is then the first order of signification (O1), language mediated by orthographic signs is the second (O2), and so on. To circumvent this technical issue, I consider language as the “primary order of signification” here called “verbalization,” the level of minimal symbolic integers like images and motifs as the “secondary 16 Indexical orders (Silverstein 2003:193-94) or orders of indexicality (Blommaert 2007:116-18) distinguish the indexicality of signs (that is, associations and connotative significance) related to their denotational use or significance as the first order of indexicality from the indexicality of social uses and users of those signs and registers in society at additional orders of indexicality. 17 Foley (for example, 1995:2 et passim) treats formulae, themes, and story patterns as different structural types of “integers” of the tradition. 432 FROG order of signification” here called “symbolic articulation,” and levels above symbolic articulation are referred to as “higher orders of signification.”18 At each order of signification, mode acts as a determinant on organizing signs in relation to syntax in the sense of “the formal relation of signs to one another” (Morris 1971 [1938]:22). In verbalization, the primary mode not only acts as a determinant on verbalization in situ: language and mode evolve in a symbiotic relationship (Foley 1999:66-83). Required alliteration, for example, not only conditions word choice in a single performance; it structures the linguistic register as a historical process, equipping the idiom to “say the same thing” within different patterns of alliteration (Frog 2015b:86-88). Language can become “entangled” with features of mode, such as formulaic expressions associated with filling certain metrical positions.19 Prominent and canonical parallelism similarly motivate the development of equivalence vocabulary, which may also evolve a high degree of formulaicity.20 The question of syntax is relevant because the basic metered frame of the mode, such as a verse, may not correspond to a basic unit of verbal utterance, which may be at the level of a couplet or stanza.21 In its turn, the primary order of signification operates as a medium for signifiers at the level of symbolic articulation. The mode of verbalization consequently acts as a determinant on the organization of those symbolic integers in relation to syntax. The signifier of a symbolic integer may be greater than a basic unit of verbal utterance, such as a series of verses,22 but the syntactic boundaries in units of verbalization nevertheless structure the metered frames in symbolic articulation. These structuring principles extend, fractal-like, from each order of signification to the next. When one order of signification operates as a mode for the next, the mode becomes one of signs rather than signals. Linguistic signs are situated in a pivotal role between the primary mode and the symbols that they mediate. Rather than symbolic integers simply becoming 18 The terms “verbalization” and “symbolic articulation” are adapted from Ruqaiya Hasan (1989:90-106; 2007:23-32), although there is a significant difference in how we each use the term “symbolic articulation.” According to Hasan (1989:98), “the stratum of symbolic articulation is where the meanings of language are turned into signs having a deeper meaning.” However, rather than considering signs mediated through language as I do here, she is concerned especially with patterns and patterning in language through a text that “provide a principle for discriminating between the crucial and the incidental” (ibid.) as well as constructing their meaningfulness. 19 The relationship between formulaic language and meter was of course the basis of Oral-Formulaic Theory (on which, see for example Lord 1960; Foley 1988; Foley and Ramey 2012; see also §2.3 below). 20 Unlike many types of formula familiar from poetries with highly structured meters, the formulaic idiom may evolve around paired equivalence vocabulary that communicate a single unit of meaning when used in parallel lines without being linked to metrical positions (Fox 2016:3-225). 21 Old Norse skaldic poetry, for example, is well known for its complex arrangements of syntax across a four-line half-stanza. There is no regular correlation between the individual line or couplet and a syntactic unit. Instead, the half-stanza unit forms a syntactic capsule and the threshold between half-stanzas forms a boundary at which metered frames of the mode and verbalization are compelled to align. 22 Some traditions allow recurrent returns in the verbalization of an image or motif in the form of multiple sequences of text interspersed with other representations. This phenomenon would require the introduction of an additional poetic system and too much elaboration to discuss here, but see for example the recurrent returns to the image of the waters into which the hero will dive, interwoven even into the description of the hero donning his armor, in the epic Beowulf (lines 1422-54). These recurrent returns through sequences of verbalization to a coherent integer in symbolic articulation are equivalent to forms of semantic parallelism where parallel members are interwoven with other verses. PARALLELISM DYNAMICS I 433 metrically entangled with verbalization as a mode, language develops connections to the symbols it mediates. The linguistic signs and symbols can then be said to index one another: they develop an association that links them in the same way that “smoke” indexes “fire.” If a performer fluent in the idiom wishes to express a particular motif, this indexicality connects the motif to phraseology that facilitates expressing it through the poetic system. Indexicality also operates in the opposite direction. The connection of particular formulae to the motif can make it transparently recognizable from the first verse before it is even described. In §3.2 below, the solitary verse jo tuli tulini koski (“already came a fiery rapids”) in a particular epic performance communicates a complex mythic image of a monstrous eagle. This indexicality can also produce connotative meanings and intertextual reference when such a formula is used in a different context, such as illustrated in Lotte Tarkka’s (this volume) analysis of a song about being a widow that engages mythological and ritual symbolism. In long epic forms prone to elaboration, for example, the verbalization may be highly variable and extended so that the signifier of a symbolic integer may be of considerable scope and only gradually apprehended.23 In a shorter poetic form like kalevalaic epic, verbalization can become a highly conventionalized complex unit that functions as a “macro-formula” for people fluent in the register (Frog 2016a:7-10 and 2016b:64-65, 76-91). The expression jo tuli tulini koski is the opening of such a macro-formula, allowing the whole image to be recognized before it is even described because the verse’s use is so specific: indexicality becomes exclusive and contextually transparent in epic narration. When indexicality becomes exclusive, the orders of signification begin to converge and apprehension of the symbol is immediate, without requiring further parsing and interpreting of language.24 In this case, the unit of language acts like a name, even if a name always remains a word and is never identical to what it refers to. Images and motifs may similarly develop indexical associations related to higher order symbols according to regular patterns of use. In Fig. 1 above, higher orders are clearly distinguished as separate layers of signification. In practice, indexicality can make the progression much fuzzier where higher order integers of the tradition are fully internalized and can be rapidly recognized. The problem can be illustrated by applying the equation O1 → O2 → O3 → etc. to the ideal model in Fig. 1, where O1 = verbalization, O2 = symbolic articulation, O3 = higher order 1, O4 = higher order 2, and so forth. If we are being told a story (O1) and hear about a girl meeting a wolf while walking to her grandmother’s through a forest (O2), we can recognize a motif that indexes the complete plot of the Little Red Ridinghood type. When this happens, the particular theme in which the motif appears may be unfamiliar to us, but we have already leapfrogged through the ideal orders of signification on the back of indexicality. For the sake of argument, let it be supposed that the theme would ideally be recognized at O3 and the plot at O4. However, the emblematic motif activates the plot already at O3 even as we may still be eagerly listening to the unfamiliar theme to unfold—also at O3. The process becomes more complicated when recognition of the plot 23 This can be a straightforward description, like the series of verses in Beowulf (lines 1422-30a) that describe the serpent-infested waters beyond which the mother of Grendel lives. Elaboration may also construct the image through references to its role in a broader narrative sequence, such as referencing what Beowulf’s armor will do during his dive as part of its description in the motif of the hero arming himself (Beowulf lines 1441b-54). 24 See also Wray (2008:17-20) on formulaic language. 434 FROG reciprocally activates other emblematic motifs and themes before they have been verbalized. The relative orders of signification retain validity and relevance in the hierarchy of relations of signs. However, internalized knowledge breaks down dependence on a strict progression through the hierarchy in order to apprehend the higher order signs. Rather than an equation of O1 → O2 → O3 → etc., the higher orders easily collapse and interact in semiosis in an equation of O1 → O2 → n. Where verbalization metonymically activates signs at multiple orders simultaneously, the equation may even reduce to O1 → n, 25 yet such activation may also be contradicted by the actual progression of performance, as signs are progressively signified through integers within the hierarchy. Indexicality allows shortcuts through the orders of signification, but the activation of signs at multiple orders in this way situates them as “enabling referent[s] of tradition” (Foley 1995:213) in relation to which the actual signs within the hierarchy may meaningfully align or contrast in the course of performance. 1.3. Formal Types of Symbolic Integer Within this approach, symbolic integers of the tradition can be distinguished according to formal type and relative order of signification. Only those relevant to the present discussion will be briefly introduced here. The order of symbolic articulation is characterized by the operation of minimal symbolic integers that can be formally distinguished in terms of “images” and “motifs.” A cultural image is approached here as a socially recognizable static symbol that corresponds to the linguistic category of a noun. It may be simple or complex (for example incorporating additional images and indexing particular motifs as part of its symbolic identity), but it is recognized as being a conventional, coherent unit. A mythic image is a cultural image of mythic quality, an emotionally invested symbol that could be engaged in mythological thinking.26 The term “motif” has been defined in various ways (see, for example, Berezkin 2015:61-62). Most commonly, it is used ambiguously for any narrative element “useful in the construction of tales” (Thompson 1955:7), often imagined as an abstract universal (see also Thompson 1955-58). A motif is here more narrowly defined as a socially conventional construction (however fixed or abstract) that engages one or more images in a relationship (“hero strikes monster”) or as undergoing change (“monster dies”). Following the linguistic metaphor, a motif is set apart from the nominal category of image by incorporating the linguistic category of a verb. As will be illustrated below, a crystallized motif may in some cases be parsed as incorporating two or more verbs. However, it is characterized as a unitary integer of the tradition in the same way that a complex linguistic sequence can form a formulaic integer in verbalization. According to this approach, both images 25 Foley (1995:96) points out that the formulaic expression in the South Slavic epic tradition, “He cried out,” toward the beginning of a performance “signif[ies] the lament of the prisoner-protagonist in the Return Song.” This one formula in context activates motif, theme, its relationship to other themes and even the narrative pattern of the epic’s plot all at once. 26 Mythological thinking is the process of thinking through mythic symbols as models for understanding the world and interpreting experience, and also as models for action and identity (see Siikala 2002:47-70; Frog 2015a: 38-39; Frog, “Parallelism Dynamics II,” this volume). PARALLELISM DYNAMICS I 435 and motifs are assumed to be tradition-dependent and/or culture-dependent even if similar integers can be found in other traditions. Themes and narrative patterns operate at higher orders of signification. Conventionally associated images, motifs, and/or equivalent sets of the same, can mediate more complex integers that become recognizable through the organized co-occurrence of their emblematic constituents. In narration, these integers are of larger textual scope and are here addressed as themes.27 Themes remain formally distinguishable as complex units that can incorporate multiple motifs and in which motifs may be repeated. 28 Themes, images, and motifs may all mediate and form a more complex signifier, here generally addressed as a narrative pattern. A narrative pattern is formally distinguishable from a theme in its capacity to organize themes into larger units within which themes may also be repeated. Narrative patterns may form further hierarchies that are not terminologically distinguished. In epic narration, for example, each episode may present a narrative pattern, as may an adventure of multiple episodes, and an epic whole of multiple adventures. Each of these may constitute a discrete traditional integer mediated through other symbols. The term “plot” is used here for a narrative pattern that is a coherent and socially recognized whole from complication to resolution, even if in practice a plot might be linked to additional narrative patterns in series. 1.4. Diagraph Analysis Du Bois presents a tool for the analysis of parallelism in co-produced conversation that proves valuable for the study of parallelism more generally. He refers to this as a “diagraph:” it is a method for organizing parallel members of groups on a grid for a visual rendering of parallelism (2007:159-62 and 2014:362-63, 376-78). This method was initially developed for the study of dialogic syntax as a tool to map the elements and structures reproduced across utterances, as in the exchange presented as example (1).29 In the terminology used here, each of the two utterances can be viewed as realizing a metered frame. The diagraph situates the second metered frame below the first and distributes the semantic lexical elements on a grid. The grid reveals Ken’s reproduction of elements from Joanne’s utterance and its grammatical parallelism; it brings out correlations of pronouns and highlights contrasts (reproduced from Du Bois 2014:377): (1) JOANNE: it ’s kind of like ↑you Ken . KEN: that ’s not at ↑all like me Joanne . 27 Like the term motif, the term theme has been used in a variety of ways and most often without clear formal criteria to distinguish it from other structural units (see for example Propp 1968 [1928]:12-13; Lord 1960:68-98; Frye 1968; Foley 1990:240-45, 279-84, 329-35). A highly conventionalized theme may crystallize into what might be described as a “macro-motif,” a predictable system or series of images and motifs comparable to a sentence of multiple clauses that operates like a complex formulaic sequence in symbolic articulation. 28 A theme is to images and motifs what a linguistic multiform (§2.3) is to words and formulae. 29 Diagraph analysis can be extended to the level of the grammar of parallel members (Du Bois 2014:388-92), which is more complex and detailed than needed here. 436 FROG Because the diagraph is used to analyse semantic parallelism, the grid will not necessarily align with metrical rhythm. In kalevalaic poetry, for example, semantic parallelism between lines normally follows a convention that “the repeated line, or lines, must not contain anything that does not have a corresponding complement [that is, word or phrase] in the first line” (Kuusi et al. 1977:66; see also Saarinen, this volume). Here, the diagraph makes ellipsis in parallel members evident when the parallel couplet in (2) is laid out in the diagraph in (3). As is conventional for kalevalaic poetry, quotations follow the orthography of (mostly nineteenth-century) transcriptions in the edition Suomen Kansan Vanhat Runot (“Old Poems of the Finnish People”) with abbreviations expanded for readability;30 standardized spellings are used for words in discussion, so minor spelling variations between these will sometimes be observed. Since lineend punctuation is an unnecessary editorial introduction of collectors or editors, I have removed it from all examples: (2) (3) Lemminkäinen, lieto31 lad Drove black grouses from a field Hens from a copse of birch (SKVR I2 742.126-28) Lemminkäini lieto poika Tempo teyriä ahosta Koppaloita koivikolta Tempo teyriä aho- sta Koppaloita koiviko- lta Drove black grouse field from hens birch.copse from The diagraph can be adapted to look at parallelism between any set of signs in a metered frame. In later sections, it will be adapted to correlate semantic and symbolic units of different scope. 2. Kalevalaic Poetry, Multiforms, and Epic Narration The majority of examples addressed here are taken from kalevalaic or Kalevala-metric epic poetry. This is a mythological epic tradition. Categories of “god” and “hero” are best described as forms of agency in this culture. Protagonists will be referred to as “heroes” for simplicity’s sake. All examples are from transcriptions of oral poetry preserved in archives. They should not be confused with the nineteenth-century national epic Kalevala, developed on the basis of collected oral kalevalaic poetry (and from which the terms “kalevalaic” and “Kalevalameter” are anachronistically derived). Most examples are quoted in short excepts taken primarily from The Song of Lemminkäinen and secondarily from The Singing Competition, both of which are briefly summarized at the end of this section. Kalevalaic epic was a short epic form. Its “ideals of tersely progressing narrative discourse differ from [those of] the ‘broadly narrative’ epics prone to delay and the elaboration of 30 In English on the issues of transcription and abbreviation of this poetry, see Saarinen (2013:36-38). 31 The alliterative epithet lieto is translated by words like “wanton,” “care-free,” or “loose.” It is found as a dialectal word for loose, tillable earth (see Frog 2010a:186-87, 193). PARALLELISM DYNAMICS I 437 detail” (Honko 1998:36). Individual epics and epic cycles varied in length and complexity (also regionally), but they can be roughly generalized as normally being between about 75-300 lines in performance. The stories of these epics were not variably realized in verse like the long epic traditions, the basis for the development of Oral-Formulaic Theory.32 Kalevalaic epics circulated as socially recognizable verbal texts in which lines and stanza-like clusters of lines were linked to particular epics in a singer’s repertoire.33 Poems were sufficiently stable in oral circulation to prompt Leea Virtanen (1968:55) to observe “a researcher can usually say without difficulty to which song particular lines belong.”34 Each stanza-like cluster of lines reflects a generative verbal system for producing “chunks” of text. Drawing on the terminology of Lauri and Anneli Honko (1995 and 1998), a generative verbal system of this type will be addressed here as a linguistic “multiform.” Multiforms are important for understanding the verbal stability of these poems and also for analysing some types of parallelism. They will be introduced further below, but when thinking about the transmission of poems, it is helpful to recognize that an episode may be constituted of only 5-10 multiforms, some of which repeat with different types of variation, while some epics may be constituted of only about 15 such multiform units.35 The degree of stability makes the poetry a valuable resource for illustrating the organization of parallelism at different orders of signification as well as interplays between orders that may be less apparent in other poetries, and thus more equivocal for discussion. 2.1. The Poetic Form and Register The Kalevala-meter is basically a syllabic, trochaic tetrameter: each line is normally eight syllables in length, although there was flexibility in the first foot (the first two positions). The meter has rules governing the placement of long and short stressed syllables (syllabic quantity rules). Verses are characterized by alliteration, although alliteration is not linked to the meter per se, and by semantic and syntactic parallelism, as well as by a tendency for longer words to be placed at the end of a line (unless this conflicts with syllabic quantity rules).36 This is a stichic 32 On the development of Oral-Formulaic Theory, see for example Foley (1988) and Foley and Ramey (2012). 33 On the application and applicability of Oral-Formulaic Theory to kalevalaic epic, see Harvilahti (1992a: 141-47, espec.); on the identification of poetic sequences with particular texts, see for example Saarinen (1994:190); Tarkka (2013:90-93) and this volume. Jukka Saarinen (1994:190) stresses that linking formulaic sequences to one epic as opposed to others should be viewed as occurring at the level of individual singers’ repertoires. This perspective accounts for certain descriptions being associated with different epics in different dialects of singing. The tendency points to a general emic conception of verbalization sequences being epic-specific. 34 Lotte Tarkka (2013:90) stresses that Virtanen’s observation concerns the song or songs with which lines were commonly associated in the tradition, but that Virtanen’s statement is problematic because the same verses could be validly used in other contexts, as Tarkka’s contribution to this volume illustrates. For discussion and illustration of stability of this poetry in oral circulation, see Frog (2016b:66-93). 35 For example, the epic known as The Singing Competition as performed by singers in the well-known Malinen family was constituted of only 14-16 multiforms in performance (Frog 2016b:72; Ontrei once concluded the epic abruptly with a variation in the penultimate multiform, M15; Ontrei’s sons Jyrki and Vassilei did not include multiforms M5 and M8 in their performances). 36 On this verse form, see further Sadeniemi (1951); Kuusi et al. (1977:62-68); Leino (1986:129-42). 438 FROG poetry, meaning it is composed in individual lines rather than couplets or stanzas of formally prescribed length. In Finnish and Karelian, an eight-position line is normally 2-4 words, of which at least two most often alliterate. The meter facilitated the crystallization of phraseology in the idiom: individual lines were inclined to evolve into formulae that were fairly stable in social circulation. Verbal conservatism nevertheless remained a social convention, which in epic exhibits an “inclination to non-variation” (Frog 2010b:99-100 and 2016b:66).37 Finnic languages lack articles, so “a” and “the” in translations lack correspondence in the original text, and articles have been omitted from translations shown in diagraphs. Finnic languages are heavily inflected. In the translations, English prepositions in most cases reflect case endings. In most cases where case endings are separated in diagraphs, they have also been translated for clarity. The verb olla (“to be”) often functions as a sort of optional particle in verses. Its present third person singular inflection on (“is”) occasionally is used as a metrical filler, although more often it may seem to have a grammatical role, as in koivuss’ on tuline kokko (“in the birch is a fiery eagle”) in examples (8) and (33), where it is in fact optional and textural. The diction of epic poetry has evolved in relation to the poetic form. The syllable-based rhythm and conventions for the placement of long and short stressed syllables were centrally accommodated by three factors: a) the flexibility of the first foot, where syllabic quantity rules did not apply, b) variable word order, which allowed words that conflicted with syllabic quantity rules to be shifted into the first foot, and c) the use of grammatical forms that affect the number of syllables to match word length to rhythm.38 Diminutive forms are found in several examples below, such as vaimo (“wife”) > vaimo-ini (“wife-DIMINUTIVE”) in sińi vaimoized vajutti (SKVR I2 703.268-69) (“so long the wife-DIMINUTIVE-PLURAL wept(?)”). Such diminutives are translated here by the adjective “dear” except where they are part of a proper name.39 Lexical extension or prolongation is also done with verbs by adding an element to the stem, such as the frequentative -lla (inflected -le-), indicating repeated or ongoing action. Such verbs can be difficult to translate when the form seems to be metrically rather than semantically driven. Lexical prolongation to match a metrical shape should be considered a formal feature emblematic of the register. Verse parallelism was fundamental to the poetry, both semantic and grammatical or morpho-syntactic. Semantic parallelism has produced uses of language that require some comment to make the examples more illustrative. Uses of words in semantic parallelism are also relevant for considering parallelism in symbolic articulation and at higher orders of signification. As noted above, each element of a parallel verse should, in principle, have a correspondent in the preceding verse, although this rule is not absolute (Steinitz 1934:157-60). Jukka Saarinen (this volume) stresses that parallel relations are between syntactic elements, so a noun may be 37 The discussion here concentrates on the general conventions of conservatism in reproduction. For additional discussion of the inclination to non-variation in other poetry, see also Frog (2011a:48-50). 38 For example, the word sampo has become widely known as the supernatural object at the center of Lönnrot’s Kalevala; this form sampo, with a two-syllable inflected stem, is actually a metrically motivated derivative of sammas, inflected sampaha- (Harva 1943:53). 38 Heroic names with a long first syllable have often been shaped with diminutive affixes to produce a foursyllable form such as Väinä-mö-inen and Ilmari-nen so that they easily produce metrically well-formed lines with the formulaic system (Frog 2016b:75-76). PARALLELISM DYNAMICS I 439 matched by a noun phrase. For example, soari (“island”) in example (6) is matched in the parallel line by niemen tutkain (“peninsula’s fringe”). Morphology may vary between parallel elements (Steinitz 1934:189-92), so soarella (“on the island”) is in the adessive case (-lla) while niemen tutkamessa (lit. “in the peninsula’s fringe”) is in the inessive case (-ssa). Such variations are not indicated in translations. Jonathan Roper (2012:91) emphasizes that “line-internal alliteration has high semantic demands, higher than those of end-rhyme.” The combination of alliteration and parallelism has produced rich equivalence vocabularies that sometimes subordinate a word’s semantics to its functional role in a semantic equivalence class. In the wake of this lexical variety to “say the same thing,” Mari Sarv (1999:126) argues that parallelism helps to “clear the alliterative haze,” or “the obscuring, changing, and extending of the conventional denotation of words when they are placed in alliterating combinations.” In the couplet “sińi itki soaren immet / sińi vaimoized vajutti” (SKVR I2 703.268-69) (“So long wept the island’s maidens / so long the dear wives wept(?)”), the words immet // vaimoiset are translated as “maidens // wives-DIM” because of their semantics in conversational speech, although both words were semantically subordinated in the poetic register as equivalents for “woman.” The verb vajottoa, on the other hand, literally means “to be boggy, slushy, squishy; to sink.” It is clearly used for alliteration, but its pairing with itkie (“to weep, lament”) is quite localized and rare, and cannot be considered generally familiar: most people would presumably have had to interpret it through itkie (or simply accepted it as parallel and thought no more about it). Parallelism between elements in kalevalaic verses may be based on a number of possible relations (Krohn 1926:80-84; Steinitz 1934:179-215). Semantic ambiguities between parallel verses are not necessarily resolved and some juxtapositions can be challenging or simply confusing. Metonymy is used as a device in representing images and motifs but not normally for producing semantically equivalent terms (but see example (10)). In many cases, tension is produced between the tendency for concreteness of language with semantic incongruity between parallel lines. The pair soari // niemen tutkain (“island // peninsula’s fringe”) produces semantic indeterminacy that can be interpreted as either, or as neither—a topographical ambiguity that participates in constructing the mythic quality of a location in the otherworld. Paired numbers like 6 // 7 produce rather than clear semantic haze (Steinitz 1934:201-03). Although emphasis here is on relationships of linguistic parallelism to symbolic articulation and parallelism between symbolic integers at different orders of signification, the operation of language in parallelism offers perspective on the types of juxtapositions that “in the given speech community act as a correspondence” (Jakobson 1987 [1956]:111). 2.2. The Problem of Symbolic Equivalence in Parallel Verses Components of an image or motif may be represented in only a single line or couplet. In many cases, a couplet or parallel series forms parallel members of groups through reference to a common image or motif. In example (21), for instance, the couplet “loati leppäseñ veneheñ / loat’i leppäseñ urohoñ” (SKVR I2 809.140-41) (“[he] built a boat of alder wood / built a hero of alder wood”) refers to two distinct elements of a coherent activity, and the veneh // uros (“boat // hero”) pair is reproduced for each of three actions in a series of three couplets. Eila Stepanova (this volume) uses the term “additive parallelism” to describe parallelism in which “the 440 FROG ‘recurrent return’ augment[s] the semantic unit with some new detail” so that “semantically parallel members are not semantically identical.” The structure of semantic parallelism in kalevalaic poetry inhibits additive parallelism at the level of verses by requiring that each element in a parallel verse correspond to an element preceding it (although see below). Verses will, then, employ syntactic and morphological parallelism along with lexical reproduction so that each verse presents a complementary unit of information. A difference is that additive parallelism involves both semantic parallelism and new information in parallel members whereas kalevalaic epic parallelism tends to do one or the other. This couplet illustrates the difference: it reproduces the formula loati leppäsen X (“built and alder-wood X”) that creates a series of acts rather than lexically varying these to produce semantically parallel expressions that refer to a single act. When attention turns from linguistic semantic equivalence to a broader view of parallelism, however, the veneh // uros (“boat // hero”) pair comes into focus as symbolically parallel elements that are reproduced through a series of couplets in (21), where they are simultaneously distinct and complementary within coherent units of narration. From that perspective, the couplets exhibit parallelism, not at the level of semantic content, but through complementary references to a coherent element at the level of symbolic articulation. The coherent element is the process of “building” magical objects, and thus this couplet presents additive parallelism, referring twice to the same activity with reference to two of its outcomes. In other cases, the boundary can blur between parallelism dependent on symbolic articulation and more general semantic parallelism. When components of an image or motif are represented by a parallel group, corresponding elements are inclined to converge in meaning or significance. In many cases, semantic parallelism remains purely linguistic: whether each element is semantically equivalent in the parallel members or they act as parallel signifiers for a common signified (for example immet // vaimoiset not as “maidens // wives-DIM” but simply “women”).40 Sometimes, however, verses follow the principle of correspondence but seem to say different things and potentially refer to different features of the same symbolic element at the next order of signification. These cases are difficult to interpret because of the possible degree of semantic flex of individual words on the one hand and the conventionalization of word pairs characterized by semantic haze or indeterminacy on the other. This problem is illustrated by the concluding couplet in example (4-5), which links different features of the body to parallel locations of magical banishment: (4) Tuonne lauloin Lemminkäisen Tuonen mustahan jokehen Manalan ikipurohon Kynsin kylmähän kivvehen [sic] Hampahin vesihakohon Thence I sang Lemminkäinen Into Death’s black river Into Manala’s eternal brook By the nails into a cold stone By the teeth into a water-log (SKVR VII1 840.155-59) 40 Compare Fox’s (2016:xi) “simple formulae based on the strict pairing of words” across lines in canonical parallelism, so that together they form a single unit of meaning. PARALLELISM DYNAMICS I (5) 441 Kynsin kylmähän kivvehen by nails into cold stone Hampahin vesi- hakohon by teeth into water- log Before proceeding, it is worth drawing attention to the fact that the distinction between an image and a motif blurs when analysed at the level of a couplet. The question of whether the symbolic integer is static (noun-equivalent) or dynamic (incorporating a verb-equivalent) becomes conflated with parsing phraseology referring to it. Here, “into Death’s black river” and “by the nails into a cold stone” are both process: they involve movement of a body (Lemminkäinen’s) into a river and of nails into a stone. The whole is dynamic, incorporating a verb-equivalent feature, and can be viewed as a motif. However, the process also results in a static state of the body being in the river and nails being in the stone—images. In this case, the verses quoted are direct speech: they refer to a past event and give an account of the hero’s location in the present—in the river—and the static image rather than the dynamic motif comes into focus. At the level of propositional information, the parallel verses seem to describe different things. Morphological and syntactic parallelism make salient a correlation between “nails” and “teeth” as parts of the hero and between a “stone” and “log” as objects or locations into which these are driven. Being stuck in features of the natural environment or dissolving into it is prominent in the epic tradition. Such dissolution is connected with powerful magic and death or the threat of death. Here, each verse can be seen as expressing a different emblematic element of a system of images/motifs that form a coherent symbolic integer, and parallelism is based on equivalence in reference to that integer rather than on semantic or analogical equivalence of the words. Within the context of the poetic system, however, what is happening in this couplet is not so straightforward. The pair kynsi // hammas (“claw, talon, nail // tooth”) is a conventional pairing in the poetry customarily for a predator’s emblematic capacities for harm (see also its use in example (22)). The pair is equivalent to the English “tooth and claw” or “tooth and nail.” This pair also appears in the description of an eagle in example (8) below, where “teeth” are anatomically inconsistent: “sep’ on hampahieh hivove / kynsiähä kitkuttauve” (SKVR I2 742.123-24.) (“that one is its teeth grinding / claws scraping). Although “teeth” and “claws” are semantically distinct, they also carry a shared meaning in combination. The pair hako // kivi (“log // stone”) is similarly established in the tradition. For example, in incantations used by a ritual specialist to “raise” his power, he may use a conventional couplet to command his power to rise “havon alta hattupiäššä / kiven alta kinnaškiäššä” (SKVR I4 11.3-4) (“from beneath the log, hat on head / from beneath the stone, mitten on hand”). In the incantations, it is contextually transparent that hako // kivi refers to a single location. The pair hattu peässä // kinnas keässä (“hat on head // mitten on hand”) seems to refer to different emblems of preparedness linked to alliterations in /ha/ and /ki/, but hat and mittens have also evolved into an established (if less pervasive) pair. In these cases, rather than the parallel word referring to the same thing as the one in the main line of a parallel group, a combination like kynsi // hammas can be seen as having a distinct meaning greater than either of its parts, a rhetorical figure known as a merism. However, in light of the poetry’s inclination to concreteness in language use, this perspective also becomes reductive. 442 FROG Verses like “by the nails into a cold stone” or “that one is its teeth grinding” seem to invoke images concretely. The hako // kivi (“log // stone”) pair has established patterns of use as referring to a single location like soari // niemen tutkain (“island // peninsula’s fringe”). In these uses, semantic difference between the hako // kivi pair collapses into a single, if hazy, image. However, there is no contextual motivation to similarly collapse pairs like kynsi // hammas (“claw, talon, nail // tooth”) or hattu peässä // kinnas keässä (“hat on head // mitten on hand”). Each element in a pair may be emblematic of a broader totality, but kynsi and hammas are both used more widely outside of the pairing than in it. The words are neither systematically paired nor is the poetic system characterized by systematic semantic equivalence. In this couplet, the most natural reception seems to be “teeth” and “nails” as distinct and complementary image elements that together carry broader implications. At the same time, “log” and “stone” are inclined to a convergence into semantic haze while reinforcing cohesion between the expressions and their semantic unity. The concreteness of the poetry generates a tension between receiving these verses as presenting distinct, but complementary, features of a symbolic unit and collapsing them into alternative alliterating phrases that are semantically equivalent. This example is interesting because one set of paired image elements seems more inclined to remain distinct while the other is more inclined to collapse in what could be considered a type of additive parallelism; together both verses refer to a single integer at the next order of representation. 2.3. Linguistic Multiforms The Honkos (1995 and 1998) coined “multiform” as a technical term to describe systems of formulaic expressions and individual words that are associated in a singer’s memory as a flexible framework for expression at the level of texture without necessarily realizing a specific unit of content or meaning.41 The Honkos’ development of this concept had a targeted aim of accounting for flexibility in the length of long epic forms that was not sufficiently accounted for by Oral-Formulaic Theory. I have developed and refined multiform theory by demonstrating that linguistic multiforms can be socially transmitted and by increasing the variety of oral poetries against which the theory is tested.42 I use a tighter definition of “formula” as a “morphemeequivalent unit” (Wray 2008:11-21) or linguistic unit forming “an integer of meaning” (Foley and Ramey 2012:80). I reject Milman Parry’s (1928:16) and Albert Lord’s (1960:4) criterion that a formula is used “under the same metrical conditions.” This criterion is circularly derived from Parry’s statistical method for identifying formulae by where they occur in verses and has no independent theoretical basis. Instead, I consider formulae that regularly or invariably occur 41 See also Honko (1995, 1998:100-16, and 2003:113-22); Frog (2010b, 2016a, and 2016b). Michael Drout (2011:447) quotes the text of the Honkos’ definition of “multiform” to refer to a(ny) traditional element’s capacity to manifest in multiple forms. Drout follows use of the word “multiform” in Oral-Formulaic Theory literature (formalized in Foley 1995:2; see Honko and Honko 1998:40-41), which is divorced from the linguistic phenomenon “with a primary locus in the mind of the narrator or singer” (Honko and Honko 1998:41). Drout’s quotation thus seems misrepresentative of the Honkos’ formalization of the term to address a more distinct phenomenon. 42 My most detailed empirical studies have been on multiforms in kalevalaic epic and incantation, and in Old Norse eddic and skaldic poetries, and with more limited data in several other traditions (see Frog 2016a and works cited there). PARALLELISM DYNAMICS I 443 “under the same metrical conditions” as an extreme on the spectrum of “metrical entanglement,” the degree to which elements of language develop conventional relationships to metrical features that organize verbalization (Frog 2016c:202-03). Rather than being merely a more or less regular set of words and formulae, multiforms may also entail lexical equivalence sets that alternate, for example to meet particular patterns of alliteration or rhyme (2009:236-42 and 2016c:163, 213-14, 216). Features of syntax may also be integrated into the system, and both lexicon and syntax may be metrically entangled (2012b:31-35 and 2016b:78-82). Increasing the range of poetry in which multiforms are explored has revealed the degree of symbiosis between multiform and poetic form (see also Foley 1999:66-83). The specific features of multiforms prove highly tradition-dependent. Multiforms also exhibit a number of formally distinct types that manifest tradition-dependent forms (Frog 2016a:7-11). Linguistically, a multiform can be a very complex generative framework. Variation is conditioned by a multiform’s formal type and, if it is recurrent within an epic performance, how it is used (Saarinen 1994:186-87; Frog 2010a:103-08 and 2016b:76-91). Within a performance, most variation in recurring multiforms is of one of three varieties. It may be morphological, such as changes in person and tense when the same multiform is used in both dialogue and thirdperson narration (Saarinen 1994:186; Frog 2016b:91; compare Urban 1991:71-72). A multiform may have one or several conventional formulae or line positions with “slots” that are completed differently in each use, for example in a dialogue: “I will give you my X [= horse / boat / sister]” (Saarinen 1994:186; Frog 2016b:85-86). Variation in whole lines can commonly occur with: a) omission of the grammatical subject or addressee, which would otherwise be the opening line or couplet (as in example (33) below); b) truncation of the last line or couplet after the first use (Frog 2016b:90-91); c) addition of a line or couplet to mark the final use in a series or poem (Saarinen 1994:187; Frog 2010b:96 and 2011b:55-56). Singer habitus varies (Kiparsky 1976:97-98; Harvilahti 1992b:87-89, 95-96), but, generally speaking, the inclination to nonvariation tends to keep variation to a minimum in epic. A recurrent multiform’s formulae tend to be reiterated verbatim and in consistent order within a performance to the degree that the alternation of one line or couplet for another is a potential indicator that the informant was having difficulty remembering the epic (Frog 2016b:89-91). Between performances by a single performer, there is some variation in the presence or absence of (mainly semantically insignificant) lines. There may be variation in line order within multiforms where this does not affect interlinear syntax or sense, and variation in the presence, absence, and number of reuses of multiforms. A singer’s phraseology of lines tends to be stable even where variation has no effect on meaning, such as saying se on (“it is”) rather than tuo on (“that is”). Semantically light phrases and verses easily vary between singers in a local or kin-group dialect. Phraseological or structural difference in multiforms between singers is an indicator of dialectal difference, although it could also be a slip, an effect of dictation on diction, idiolectal, or attributable to some unknown factor. In kalevalaic poetry, multiforms crystallize around minimal content units. Multiforms in the Tulu long epic form analysed by the Honkos (1998:52-55) allowed one multiform to be embedded into another, extending or prolonging and elaborating it. In contrast, kalevalaic epic multiforms generally do not allow embedding or interweaving (although see Frog 2016b:77, 88). They are simply abutted into chain-like series, normally without connecting tissue. Multiforms 444 FROG construct metered frames for integers of symbolic articulation, for which they can operate as macro-formulae. When a motif, theme or larger sequence of narration is transferred into a different narrative context, the respective multiforms are transferred with it.43 The verbal crystallization of multiforms also makes parallelism between recurrent symbolic integers salient. Socially circulating multiforms present unambiguous units of composition. Dyadic structuring is very common in kalevalaic epic multiforms. Whereas the potential symbolic parallelism in couplets illustrated in example (4) was ambiguous, parallelism in dyadic multiforms clearly presents complementary units of information about a symbolic integer of narration. Verse parallelism normally manifests in presenting each of the multiform’s two elements. Clear, two-element structures may raise the question: why not analyse the two parts separately? The dyadic constituents can be either static, like images, or dynamic, like motifs, so it could seem natural to isolate them as units. However, motifs and images are here approached in emic terms, and multiforms provide a methodological basis for identifying what constitute unitary symbolic integers in narration (a precondition for analysing parallelism based on those units). Where two elements form parts of a consistent multiform, and especially where they are not realized independently of that multiform, they can be considered to form a coherent integer of the tradition. Rather than an image or motif being constituted of multiple images and motifs, it should simply be considered complex in the same way that a formulaic sequence is complex as an integer of language.44 Established use as only part of a multiform, rather than as an independent unit, is an indicator that the particular element of narration is not an independent integer of the tradition. The way multiforms work in kalevalaic epic differs from how they operate in many other types of discourse. However, they facilitate discussion and illustrate varieties of parallelism above the level of verbalization and thus illustrate a framework that may be adapted to more variable forms of verbal art. 2.4. Kalevalaic Epics from which Examples Are Taken The Song of Lemminkäinen was arguably the most popular epic in the period of documentation. It is preserved in around 400 variants and fragments,45 mostly in hand transcriptions but some also recorded in audio. The stable core of the epic was the hero’s dangerous journey from his home to an otherworld community. The name of the hero is commonly Lemminkäinen, his destination Päivölä (“Sun-LOCATION”), and the festivity identified 43 The process of the renewal of a multiform for a particular symbolic integer also occurs, but this process has not yet received detailed investigation. 44 For example, a kalevalaic formula like sanan virkko, noin nimesi (“said a word, thus uttered”) can be broken down into four lexical items. However, the whole phrase remains a linguistic integer in the register of kalevalaic epic: it simply means “said” while forming a complete verse. This formula might look like sanan virkko (“said a word”) and noin nimesi (“thus uttered”) are two separate formulae, but a cursory search of the SKVR digital database does not support this: there are some minor lexical variations and I observed one instance where the second half-line was completed differently, but neither phrase seems to have circulated socially as an independent formula. 45 For an overview of the several hundred examples published in SKVR, see Frog (2010a:72-98). PARALLELISM DYNAMICS I 445 in parallel lines as a jumalisten juominki (“drinking-feast of the gods”). Dialectal variation of names is not relevant here. The epic customarily begins with the image of a fire that is connected with beer-brewing. Invitations to a wedding or feast are sent to everyone except the protagonist. The hero determines to crash the party and asks his mother to fetch his armor. A dialogue ensues in which his mother forbids him from making the journey and warns him about three “deaths” on the road. Her warnings unheeded, she brings the hero his armor. The hero’s journey follows: he passes each of the three dangers. When he arrives at the feast, he is received with hostility and ostensible, but life-threatening, hospitality. The plot then has two distinct branches that may also be linked in series. In one, the hero is slain by magic and ends up in the river of death. His mother learns of his death from an omen and travels to Päivölä in order to discover what has happened. She then travels to the river of death, dredges her son out with a magic rake, and (more often than not) successfully resurrects him. In the other branch, the hero enters into a duel of magic with the host, followed by a duel of swords. He wins both and the host is killed. The hero returns home and asks his mother where he should hide from the anticipated retribution. He then sails to an island of women, where he sleeps with all of the women and, finally, leaves. The epic concludes with the departure of his ship. A few examples are taken from The Singing Competition, in which a young hero, normally called Joukahainen, drives his sleigh against that of the demiurge Väinämöinen and they crash together. They have a competition of knowledge to determine who should give way to the other. Joukahainen claims participation in the creation of the world, which Väinämöinen identifies as a lie and magically “sings” Joukahainen, causing him to dissolve into the landscape. Joukahainen offers his most valuable possessions to ransom his life, but only when he offers his sister does Väinämöinen accept and Joukahainen returns home to tell his mother. Parallelism across episodes of the Sampo-Cycle are also discussed in §5.3, but the relevant information about the plot will be introduced in conjunction with the comparison. 3. Image-Based Parallelism in Multiforms Units of narration in kalevalaic epic were characterized by mythic images. In examples like (4) above, mythic images can be integrated elements of motifs and also implicit as the static outcome of a motif. A mythic image that itself constitutes a complete integer of narration will exhibit a devoted multiform. In this section, the focus is on parallelism that is based on images in symbolic articulation. Focus is narrowly on parallelism within a multiform, of which two or more elements make reference to a coherent mythic image. Parallelism between images also occurs at the level of whole multiforms, but, in epic, image-based multiforms do not generally appear juxtaposed in series. Parallelism of images occurs in sequential motif-based multiforms, as in examples (26-27). Image-based multiforms may also be parallel elements in higher order integers, as in the example of theme parallelism in (33). Image-based parallelism is readily observable in multiforms with a dyadic structure that can be regarded as a type of parallel construction. Example (6) presents parallelism between references to the same image in positive and negative terms. Example (8) illustrates how the two 446 FROG parts of a dyadic multiform may present complementary information about the same mythic image. Example (10) presents an example of contrastive image-based parallelism without negation. 3.1. Parallelism through Negation A common form of contrastive parallelism in kalevalaic epic is expressed through negation. Contrastive parallelism is characterized by forming “concurrence of equivalence on one [. . .] level with disagreement on another level” of expression (Jakobson 1981 [1966]:133) so that the contrast rather than the equivalence is highlighted. Contrastive parallelism in this poetry tends to be formally more complex than semantic verse parallelism of the A1 A2 type. Contrast rather than equivalence is indicated by prefixing ‘-’ to a parallel member, such as A1 -A2. Rather than two or more verses that “say the same thing,” contrastive parallelism requires verses that say different things, each of which may be expressed in two or more semantically parallel verses. Thus, rather than a simple structure of an A1 -A2 type, contrast occurs across groups in a structural hierarchy in which contrasting members are each expressed by a semantically parallel group, such as (A1 A2)1 -(B1 B2)2. Although phonic, morphological and syntactic parallelisms may produce cohesion across verses in contrastive parallelism (Jakobson 1981 [1960]:41), this type of parallelism does not operate purely at the level of linguistic signs: it depends on referring to the same symbolic integer in different ways or on creating a contrast between such integers. A conventional opening to The Song of Lemminkäinen provides a simple example of parallelism in positive and negative terms referring to the same mythic image. The mythic image is rendered through a devoted, crystallized multiform that has been recorded in over 150 variants with remarkable consistency. In this case, two semantically parallel couplets are interwoven by resonant alliteration that extends across lines and couplets (underlined). The two couplets are alternative references to the same mythic image. Together, they form a complex, hierarchically organized parallelism in an (A1 A2)1 -(B1 B2)2 structure: (6) (7) Šavu šoarella palavi Tuli ńiemen tutkamešša Pieńi ois šovan šavukši Šuuri paimośen palokši Smoke on the island burns Fire on the peninsula’s fringe Small it would be for the smoke of war Great for the blaze of a shepherd (SKVR I2 771.1-4) Šavu šooare- lla Tuli ńiemenn tutkame- šša Pieńi Šuurii ois šovan paimośenn paalavi Smokke island on Fire peninssula’s fringe on šavuu- kši Small paloo- Great kši wouuld be buurns war’s smokke for shepherd’s blazze for Variation in this multiform primarily occurs in: a) the line order of the second couplet or the second couplet’s omission (although in some cases apparent omission may simply reflect PARALLELISM DYNAMICS I 447 transcription practices); and b) occasional transposition of pieni (“small”) and suuri (“great”) or lexical variation between metrically and semantically equivalent terms such as tuli (“fire”) in the place of palo (“blaze”), and so on. Both couplets also frequently exhibit ellipsis of the verb in parallel lines. In his preliminary survey of what he describes as negative parallelism in kalevalaic poetry, Felix Oinas (1985 [1976]:78, especially) observed that it follows a characteristic schema as a complex rhetorical figure of three parts, any of which could be presented by a single line or a series of parallel lines. He considered that the first part could be omitted but that the second and third appeared necessary: a) the introduction of an element; b) a negative antithesis or negative analogy; c) a new positive solution. The complex figure addressed by Oinas is frequently realized as a coherent multiform (see also examples (26-27) below). The multiform in example (6) presents a positive element in the first couplet followed by a negative comparison in the second, describing the fire in terms of what it is not, and this is most frequently followed by a positive statement: X olutta keitti (“X was brewing beer”) (Frog 2010a:372). Within the poetic syntax, the juxtaposition of two elements is often an indicator of a causal relation. In this case, the juxtaposition of an agent brewing beer with the mythic image of the fire yields the interpretation that the fire is for brewing the beer. Together, these realize Oinas’ three-part schema. The formula X olutta keitti could appear as a discrete line but was predominantly used as the onset of another multiform describing the beer-brewing. A short opening episode to the epic would often be formed with the fire-multiform, the beer-brewing, and the sending out of invitations to a feast to everyone except the hero. The formula X olutta keitti may be absent and another multiform introducing the motif of beer-brewing could appear in its place, and the beerbrewing could also be elaborated with a full narrative pattern or plot of The Origin of Beer. In the variant quoted above, more than 50 lines are devoted to the beer-brewing (SKVR I2 771.5-59). In any case, rather than the figure of negative parallelism as a coherent multiform, the three-part structure produces cohesion between the mythic image in (6) and a subsequent multiform or episode of beer-brewing. Oinas’s full schema is here fully realized only at a compositional level above the multiform. The second couplet in (6) implies an uncertainty that suggests the fire is perceived from a distance. The juxtaposition of the multiform with a description of the agent present at the fire involves a shift in perspective that could be described as a change in scene.46 On the other hand, the multiform in (6) could also be used independently: the positive solution of Oinas’ schema was not required (for example SKVR II 196.1-4). Multiforms with a dyadic structure always appear to realize a coherent unit of narrative content, but this does not prevent multiforms from 46 Dell Hymes (1981:170-75) uses “scene” as a formal, structural unit of narration between what he calls a “stanza” (164-70) and an “act” (175). In the narratives that Hymes examines, as in kalevalaic epic, a change in scene has less to do with location than with participants in narration and their relations (171). In relation to the terms used here, Hymes’ stanza would equate to a verbalized multiform. A “scene” could be comprised of one or several themes organized through a narrative pattern. An “act” would be a narrative pattern that forms an episode in the plot. There is a fundamental difference between Hymes’ approach and the approach used here. The formal units beyond the stanza described by Hymes concern the structure of narration at the level of verbalization. The formal units beyond the multiform described here concern meaning-bearing integers of the tradition that are mediated verbally. The approaches are fully compatible, they simply focus on different things. 448 FROG being situated in relation to one another through larger parallel structures in the progression of narration. 3.2. Complementary or Additive Image-Based Parallelism It is not uncommon that dyadic multiforms are structured so that the two elements give complementary information about a common mythic image or motif. This can be compared to additive parallelism. Each element of the multiform introduces the same image but includes complementary information. The units of information do not necessarily overlap except insofar as they refer to a coherent symbolic integer. The parallelism may therefore be seen as complementary rather than additive in the sense that it both reproduces previous information and augments it with new information. An example is the description of a fiery eagle as a “death” to be passed by the hero on his journey to the otherworld.47 In this multiform, the first element is normally characterized by a complex parallel structure. The opening line presents the site of the image followed by a series of syntactically parallel lines, each narrowing the field of focus with the introduction of an additional image element. The progression is built on anaphora, repeating the last word as the first word in the following line, until reaching the eagle.48 The second part of the multiform complements this image with characterization of the bird in a parallel couplet accompanied by a line expressing its intentions towards the hero: (8) Jo tuli tulini koski Kosell’ on tulini korko Koroll’ on tulini koivu Koivuss’ on tulini kokko Sep’ on hampahieh hivove Kynsiähä kitkuttauve Peän varalla Lemminkäisen Already came a fiery rapids On the rapids is a fiery shoal On the shoal is a fiery birch In the birch is a fiery eagle That one is its teeth grinding Its claws scraping Ready for the head of Lemminkäinen (SKVR I2 742.119-25) 47 On this danger, see Siikala (2002:310-14) and Frog (2010a:380-81); on the multiform, see Frog (2016b: 84-85, 89-90). 48 This structuring of verses was described by Kaarle Krohn (1918 I:79) as “chain-type” (ketjuntapainen), by Steinitz (1934:120-22) as “chain verse” (Kettenverse), and by Robert Austerlitz (1958:63-69) as a “terrace” structure. PARALLELISM DYNAMICS I Jo (9) Se- tuli tulinni koski Alreeady came fiery rapids Kose- lll’ on tulinni korko rapids on is fiery shoal Koro- lll’ on tulinni koivu shoal on is fiery birch Koivuu- sss’ on tulinni kokko bircch in is fiery eagle p’† on haampahhieh Kynsiäähä † 449 hivove kittkuttauve It- pä† is teeeth grinding Claws scraping -pa/-pä is an emphatic particle, hence “it”-pä is translated above as “that one.” The quotation above is taken from the hero’s journey. It is preceded by the same multiform’s use in the hero’s dialogue with his mother warning him about the “deaths” he will meet (SKVR I2 742.69-75; see also §5.2 below). There are only two variations between reproductions of this multiform. The opening formula jo tuli tulini X (“already came a fiery X”), used both for this and the other “fiery” dangers, appears in the dialogue as niin tolou tulini X (“so will come a fiery X”). The word koski, here inflected kosell’ (“on/at the rapids”), was inflected kosess’ (“in the rapids”) in the dialogue. The first part of this dyadic pair is the most historically stable in social circulation. It exhibits slight variation in the phraseology of the opening line and in the number of elements in the parallel series. This series also exhibits minor lexical variation in different dialects, such as between korko (“shoal”) and luoto (“skerry”) (Frog 2016b:85; see example (33) below). The second element of the multiform is less consistent and could be omitted entirely. Omitting the second element without compromising narrative coherence is comparable to omitting the second of two semantically parallel lines. The multiform is used to represent a symbolic integer of narration through language, and we thus focus on language in order to access the symbolic integer. Within the tradition, however, the multiform is a macro-formula for that integer—a complex “word” (Foley 1999:67-70) of the registral lexicon. As such, the macro-formula communicates the image as a complex signifier with a scope of several verses. People fluent in the register need only recognize the multiform to activate the image without necessarily reflecting on the propositional information of each line (see also Wray 2008:17-20). The linguistic representation may be thought of as a sort of crystallized collage of image elements and rhetorical devices while, in practice, the symbolic integer will be activated already with the first line. In fact, this multiform can be found collapsed into a single line in performance, such as jo tuli tulini koski (SKVR I2 805.120), mentioning the eagle only in the following multiform to describe how the hero overcomes the threat (805.123). Reducing a multiform may not seem like good storytelling, but, for someone fluent in the register, there is absolutely no ambiguity that the whole mythic image has been introduced—at least where the particular formulaic line is exclusive to a macro-formula multiform. The potential for reduction makes it more interesting to observe that omission of the second element of the dyadic structure is a scattered phenomenon in the tradition and did not lead the first element to be used socially as a simplex multiform. Although the second element exhibits a variety of variation by dialect, the parallel structure was historically maintained. 450 FROG 3.3. Contrastive Parallelism of Opposed Images The preceding examples of dyadic multiforms offered examples of parallelism in which each of the two elements referred to the same image. Image-based parallelism can also involve the juxtaposition of two images with negation (Jakobson 1987 [1960]:41) or without. The following example is taken from the hero’s sword duel with the host of the otherworld feast. In the dialogue leading up to the killing, the hero proposes that the duel be taken outside. His proposal is formulated in terms of parallel images representing alternative outcomes of the duel in different locations. Each alternative is represented by a couplet of semantically parallel lines. In order to bring out certain features of the verses connected to the forms of words, glosses are included in the diagraph (COMP = “comparative;” FREQ = “frequentative;” 3rd.PL = “third person plural”): (10) (11) Pihall’ oñ veret paremmat Kagaroill’ on kaunehemmat Pessüt penkit hierelömmä Hüväñ tuvañ turmelomma Pihaa- ll’ oñ verret Kagarooi-- ll’ on In the yard is blood better In the manure is more beautiful Scrubbed benches we would spoil The good house we would ruin (SKVR I2 704.187-90) paare- mmat kaunnehe- mmat Pessütt penkit hieere- lö- mmä Hüväññ tuvvañ turrme- lö- mma Sem manttic diiagraaphh of pparaallel coouplets: Yard better- COMP beautiful- COMP in iss bloodd Manuree in iss Scrubbeed bennchhes spooillGood hoouse ruiin-- FREQ- 3rd.PL FREQ- 3rd.PL [Dirty] ouutsidde bloood possitive (adjective) Clean innsidee negative (vverb) These couplets are very tightly connected at several levels. Patterns of phonic parallelism are prominent, reinforcing cohesion of the multiform as a unit. Each couplet begins with alliteration in /p/. The grammatical rhymes of the comparative plural -mma-t in the first couplet rhyme with those of the frequentative third person plural -lö-mmä / -lo-mma in the second.49 The cohesion produced by phonic parallelism is an indicator that the juxtaposition of these couplets should be interpreted as a syntactic relation. Comparative adjectives in the first couplet suggest that the relation is comparative rather than causal. The two couplets are grammatically quite different, but they have an equivalent rhythmic-syllabic organization of semantic units in the meter. This organization is reinforced by structural correspondence. In the first line, the rhythmic-syllabic structure is 3-2-3 (representing the number of syllables in each word) or 49 In line 188, -hem- in kaunehemmat would be metrically stressed and (probably) rhythmically prolonged (on Viena region singing, see Kallio, this volume). The prominence of this syllable could potentially be perceivable as alliterating with hierelömmä and hyvän in the subsequent lines (Frog and Stepanova 2011:201). PARALLELISM DYNAMICS I 451 2-1-2-3, depending on whether pihall’on (“in the yard is”) is viewed as a three-syllable unit (giving a 3-2-3 structure) or on (“is”) is treated as a separate word rather than a variation in phonic contour (giving a 2-1-2-3 structure). With the ellipsis of “blood” (veret), the second line has a 4-4 structure or a 3-1-4 structure, distributing semantic units as half-line units. The following couplet has a repeating 2-2-4 structure, with a corresponding distribution of semantic units between noun phrases (adjective + noun) and verbs. These features of sameness make the contrasts between semantic units more salient, which are brought to the surface in the semantic diagraph. Outside, and the dirtiness that is implicit in the manure, is contrasted with inside represented as clean and good. The second couplet reinforces the message through the contrast of adjectival positive valuation and implicit negative valuation in verbs of destruction. In addition, the nouns and noun phrases exhibit a chiastic structure across the couplets of {PLACE}—{THING} —{THING}—{PLACE}. Contextually, “blood in the yard” transparently refers to the inevitable spilling of blood in the duel. Notably, the verses focus not on the action but on its outcome. The outcome represents an image of blood in a location as a symbolic integer. For someone unfamiliar with the tradition, the second couplet might be interpreted as making a mess of the house by having a big fight. However, the duelling referred to is highly ritualized: basically, the duellers stand facing each other and each takes one swing at his opponent by turns. 50 Thus, a concern for “spoiling scrubbed benches” does not reflect a worry that combatants will be bumping into things. Just as “blood” undergoes ellipsis between lines of the first couplet, it also undergoes ellipsis as an image element between the parallel couplets. The concern for disruption of the space allusively refers to an image of blood transposed into that location. Although it is possible to analyse many aspects of parallelism between these couplets at the level of verbalization, the contrast is best approached as setting images in opposition to one another at the level of symbolic articulation. 4. Motif-Based Parallelism As with images, multiforms in kalevalaic epic crystallize around motifs as integers of narration. Where such multiforms have a dyadic or more complex structure, they can have internal parallelism like those in examples of image-based parallelism shown above. In this case, two or more elements are placed in relation to one another to realize the conventional motif. For motifs, the second element of the dyad will present additive information and the juxtaposition will often be interpreted as causally related to the first. The juxtaposition forms a parallelism by presenting complementary elements of, or information about, a single integer in symbolic articulation. Examples (12) and (15) illustrate motif-based parallelism within the realization of a multiform. Examples (17-19) reveal the potential for the dyadic structure of a multiform, and thus its parallelism, to dissolve in variation. Examples (21-22) illustrate parallelism in multiforms with more than two elements. 50 The structure of the duel is simply built into narration and should not be considered a social reality of the period of collection. This type of duel is probably one of the many elements of the epic tradition that is historically rooted in an Iron-Age milieu. It corresponds to the duelling tradition of Iron Age Scandinavia that was rapidly outlawed with the spread of Christianity. 452 FROG Motif parallelism also occurs across juxtaposed multiforms. The interface of the motif as a symbolic integer with a multiform as a linguistic system for its representation produces what Greg Urban (1986:26-29; 1991:79-104) refers to as “macro-parallelism,” or parallelism based on reproductions of multi-line text sequences. Sameness in the reproduction of these sequences has the potential to foreground variations between them. In Urban’s usage, linguistically mediated signs are reduced to semantic and structural sequences of language, and therefore also fall under his use of macro-parallelism. Here, macro-parallelism is addressed in the specific sense of linguistic macro-parallelism, which manifests saliently through the reproduction of multiforms that may vary morphologically and/or with open slots, reproducing a common framework. Examples (24-25) and (26-28) illustrate motif parallelism across multiforms and macroparallelism. Nevertheless, macro-parallelism as a phenomenon can in principle manifest at any level in signification. Viewed in this way, macro-parallelism can be described as parallelism between stretches of discourse that markedly exceed the metered frames of that level of signification. As will become apparent, macro-parallelism at one order of representation commonly appears to be both a symptom and an outcome of parallelism at the next order of signification. In other words, linguistic macro-parallelism results from motif parallelism, rather than being independent of it (see also §5.3). 4.1. Motif-Based Parallelism in Dyadic Multiforms Dyadic multiforms with internal motif-based parallelism tend to be organized as two groupings of lines, each built on some form of parallelism, even if not all lines are members of parallel groups. The interweaving of complementary elements through the repetition of words and structures tightens a multiform as a verbal system and seems to support its stability in social circulation. When I was looking at hierarchies of complexities in parallelism between verses and couplets, a remarkably stable yet versatile multiform led me to attend to parallelism in the images and motifs that it mediates. This multiform is from The Singing Competition, where it describes two sleighs colliding and becoming entangled. It is organized by first using metonyms for the sleighs to describe them getting stuck together, and then describing liquid running from (some of) the same metonyms as an emblem of the straining of the horses. This multiform is additionally interesting because the motif concerns a transition from movement to stasis. The dynamic quality of the motif—the process of change—manifests in the flow of performance as the transition from a description of the two heroes driving their sleighs toward one another. At the same time, the process of the motif produces a static state of two entangled sleighs with straining horses, in which these act as parallel images:51 51 The ambiguity between dynamic and static representation is also present linguistically, where the verb puuttui can be interpreted as either “became stuck” or “were in the state of being stuck,” and the verb vuoti can similarly be interpreted as either “leaked, ran, flowed” or “was leaking, running, flowing.” PARALLELISM DYNAMICS I (12) (13) Puuttu vemmel vembelehe Rahis rahkehen nenähä Veri vuodi vembelestä Razva rahkehen nenästä Stuck shaft-bow to shaft-bow Trace to trace’s end Blood flowed from the shaft-bow Fat from the trace’s end (SKVR I1 153a.4-7) Puuttu vemmel vem mbeleRahis Veri he rahhkehe- n nenä- vuodi vem mbele- Razva 453 rahhkehe- n nenä- Stuck shaft.boow shaft.boow hä Tracee stä Blood flowed shaaft.bow stä Fat trace trace to ’s end to froom ’s endd froom The semantic parallelism between the lines of each couplet is augmented with lexical reproduction and features of grammatical parallelism between the couplets shown in (14):52 (14) Puuttu vemmel vempele- he / Rahis rahkehe- n nenä- hä vuodi *Veri vempele- stä / Razva rahkehe- n nenä- stä Stuck shhaft.bow shaft.bow to / Trace trace ’s end to flowed shaft.bow from / Fat trace ’s end from m *Blood * This word has been moved from its position in the verse to make the parallelism more visible. The parallelism between the couplets reveals their tightly knit correspondence. They vary in: a) the verb in the first line, b) in the parallel terms for the grammatical subject, and c) the inflections at the ends of each couplet’s lines (morphological variation). It is characteristic of this multiform that the verb and the metonym for sleigh in the first couplet are metrically equivalent (two-syllable words with a long first syllable) and can vary in order. This is not the case in the second couplet, where most terms for liquid like veri (“blood”) have a short first syllable, which inhibits their use in the second foot for metrical reasons. Lexical reproduction and grammatical parallels reinforce the coherence and cohesion of this multiform as expressing a single unit of narrative content. The prominence of grammatical parallelism across these couplets is affected by variations in different dialects. The second element is also completely omitted in some dialects: the symbolic integer in narration is communicated unambiguously without it. The strategies for reinforcing coherence and cohesion are otherwise found across dialects from the White Sea to the Gulf of Finland.53 Throughout these singing regions, the dyadic elements of the multiform operate as parallel members of a group through common reference to the same symbolic integer, or the symbolic integer is rendered through the first metonymic element only. 52 The diagraph in (14) has previously appeared in Frog (2016b:80). 53 For a survey based on 135 examples, see Frog (2016b:78-82). 454 FROG In the preceding example, semantic parallelism is the basis for verses in each element of the multiform, and this is common. However, there are also other varieties of parallelism like the chain or terrace structure describing the fiery eagle in example (8). These strategies also provide means of extending forms of parallelism and reproduction in verbalization across the elements of a multiform, reinforcing its cohesion as a distinct unit. This can be illustrated by a response of the hero to the eagle of example (8): (15) (16) Laulo leppäseñ venehen Laulo leppäsen isännäñ Melañ leppäsen kätehe Koški koppasi veneheñ Kokko koppasi isännäñ [He] sang a boat of alder wood Sang a master of alder wood An oar of alder wood in his hand The rapids seized the boat The eagle seized the master (SKVR I2 811.185-89) Laulo leppäseñ venehen Sang alder.wood boat Laulo leppäsen isännäñ Sang alder.wood master leppäsen *Melañ alder.wood *Oar käte- he hand in Koški koppasi veneeheñ Rapidds seized booat Kokkoo koppasi isännnäñ Eaglee seized maaster * An asterisk indicates that the word appears in a different order in the text. The motif as a narrative unit is the feat by which the hero passes the danger. Parallelism occurs in the two commensurate, juxtaposed elements of narration that together form a distinct unit. The verse structure can be described as (A1 A2 *A3)1 (B1 B2)2, where the third line (*A3) varies from the preceding two while still being perceived as parallel. In this case, the dyadic structure is unambiguous because of the change in grammatical subjects and the verb. Cohesion in each parallel group is increased by the reproduction of words in the same metrical positions in each sequence: leppäni (“alder wood”) and lauloi (“sang”) in the first; koppasi (“seized”) in the second.54 The dyadic structure is reinforced by the corresponding ordered reproduction of veneh / isäntä (“boat / master”) in the same metrical positions in parallel lines. Parallelism in the syllabic structure of verses (2-3-3) is found through the whole multiform. The co-occurrence of all of these features provides the unit with strong cohesion. The lexical reproduction of koski (“rapids”) and kokko (“eagle”) link this multiform to the preceding image where it is used in a description of the danger such as (8), and to which this is the hero’s response. Linguistic parallelism and reproduction thus has the potential to produce cohesion between elements of a multiform and to create links between multiforms. 54 Krohn (1918 I:79) called this “catalogue-type” parallelism (luettelontapainen). See also Austerlitz (1958:45-51). PARALLELISM DYNAMICS I 455 4.2. Breakdowns of Motif-Based Parallelism Not all multiforms have a regular or socially stable dyadic structure. Variation (for example by dialect) can also cause a dyadic structure to lose salience, or the second element may be omitted. In The Song of Lemminkäinen, the hero’s responses to dangers on the journey vary regionally and locally more than the dangers themselves. Some of the responses are also transferred from one danger to another. The multiform reflected in (15) is one of several types of response to the eagle and exhibits several types of variation. Variations include an adaptation for passing a different danger in which the alder-man appears with a horse rather than a boat and the second element is changed (SKVR I2 784.49-53).55 Variations in which parallelism is impacted by changes in rhythms of verbalization are more significant in the present context. The examples of dyadic multiforms quoted above show that the parallel members of the dyadic structure are commensurate in scope as paired groups of parallel verses. Changes in the rhythm of presentation change the metered frames of the elements narrated, as in the following example from the hero’s dialogue with his mother: (17) Loajiñ leppäseñ veneheñ Loajiñ leppäsen isännäñ It̮e istuvuñ jälellä Kokko koppasi isännäñ I will build a boat of alder wood I will build a master of alder wood Myself I will sit in its wake The eagle seized the master (SKVR I2 806.95-98) The first parallel couplet forms a clear unit, and the third line is easily perceived as forming a parallelism with that couplet, adding an element of information to the unit of action. These three verses thus produce a hierarchically structured parallelism of (A1 A2)1 (B1)2. The fourth line presents a unit with a new grammatical subject. The verse as a linguistic unit is equivalent in formal scope to the preceding verse. However, it is neither in full semantic parallelism with that verse for an (A1 A2)1 (B1 B2)2 parallelism, nor does it present a third member to the parallel group for an (A1 A2)1 (B1)2 (C1)3 parallelism. The fourth line is a metonym of the same motif, but the element’s counterpart is the preceding three lines which formally differ in both scope and complexity. The pattern can be described as ((A1 A2)1 (B1)2)1 (C1)2. The perception of parallelism is unlikely to manifest because the metered frames are not commensurate: the fourth verse merely follows a complex, multi-verse unit. A multiform’s dyadic structure may also be blurred through variation in syntax. In some cases, the dyadic elements can be clearly presented with reproduction of the verb or a different verb in each. Where both elements involve the same grammatical subject, they may also reduce to a single subject and verb. The basic pattern of the hero’s action and its relationship to the danger remain, but the change in syntax affects the organization of information. This can be seen in the comparison of two variations of a different response to the fiery eagle. Example (18) presents a dyadic structure in which the grammatical subject is reproduced with the introduction 55 Use of a different second element can obscure the alder-man’s significance at the level of verbalization (SKVR I2 774.251-56), but the motif as a symbolic integer would still be clear to anyone competent in the tradition. 456 FROG of a new verb, syntactically marking a caesura-like divide in the metered frame. In example (19), the grammatical subject is introduced in a full line, the core of the multiform in three lines, and the unit concludes it with a couplet situating the danger relative to the hero: (18) Siitä hän tempoi teyriä ahosta Koppeloita koivikolta Ne hän työnsi syöjän suuhun Partahan palan purian Leukahan lesottajan Then he drove black grouses from a field Hens from birch copses Those he shoved into the eater’s mouth Beat into the beard of the biter Into the jaw of the swift one (SKVR I2 726.130-34) (19) Lemminkäini lieto poika Tempo teyriä ahosta Koppaloita koivikolta Syöksi syöjällä kitaha Peässä iellä matkamiestä Lemminkäistä liijatenki Lemminkäinen lieto lad Drove black grouses from a field Hens from birch copses For eating into the eater’s maw Ahead, in front of the travelling man Of Lemminkäinen oh indeed (SKVR I2 742.126-31) The difference between these examples is made more apparent through a semantic diagraph, where a double line marks the dyadic division in the example of (18): (20) Lemminkäinen drove birds2 drove birds2 shoved into danger3 into danger in front of Lemminkäinen2 * Superscript numbers indicate the number of reiterations of semantic units in parallel lines. The dyadic elements in example (18) are observable in example (19), where they could be received as couplet parallelism forming an (A1 A2)1 (B1)2 parallelism. However, the final couplet situates the line syöksi syöjällä kitaha (“for eating into the eater’s maw”) in the middle of a larger syntactic sequence. As a couplet, this conclusion disrupts the salience of the preceding parallelism because of its own structural resonance with the preceding couplet. The multiform is realized with a symmetrical rhythm of orphan-line + couplet + orphan-line + couplet. Rather than supporting the connection between syöksi syöjällä kitaha and the preceding couplet, this structure inclines to either grouping it with the final couplet or linking it to the opening line. In either case, the structure is counter to the parallelism in (18). In addition, the final couplet is recurrent, giving it emphasis over the preceding line. Basically, everything in (19) works against a salient perception of parallelism as a dyadic multiform. PARALLELISM DYNAMICS I 457 4.3. Extended Motif-Based Parallelism Kalevalaic multiforms are not limited to a two-part complexity, and they can also exhibit parallelism betweem more than two elements. In example (17) above, parallelism was not perceivable because itse istovun jälellä (“myself I will sit in its wake”) created an ((A1 A2)1 (B1)2 C1 structure. The difference from the (A1 A2 *A3)1 (B1 B2)2 structure in (15) is not very great, but it proves significant. An additional element can also be introduced between the first couplet and the eagle’s response that participates in parallelism and reinforces it by prolonging the series (A1 A2)1 (B1 B2)2 (C1 C2)3 structure, as in (20): (21) Loati leppäseñ veneheñ Loat’i leppäseñ urohoñ Tüöñsi leppäsen veneheñ Tüöñsi leppäsen urohon Koški koppasi veneheń Kokko koppasi urohoñ [He] built a boat of alder wood Built a hero of alder wood Pushed the boat of alder wood Pushed the hero of alder wood The rapids seized the boat The eagle seized the hero (SKVR I2 809.140-45) This three-part series is systematically structured to make a tightly organized textual unit. Lexical reproductions within each couplet unite syntactically and morphologically parallel lines while lexical reproductions of the veneh // uros (“boat // hero”) pair across all three couplets unite them as three parts of a unified series. Rather than two elements, this variation on the social multiform presents a series of three, parallel, metonymic elements of the motif as an integer in symbolic articulation. In kalevalaic epic, multiforms seem to remain around 2-8 lines in use, but longer examples can be found. The following example is taken from The Singing Competition. It describes Väinämöinen “singing” his adversary into the landscape. In this dialect, the multiform developed a complex four-element form. Each element has a two-part structure. The first part uses the formula laulo X (“sang X”) in which X is the object of the verb—the person or thing that is magically “sung” into a new state. The second part is a line or series of lines telling what the object is sung “into”: (22) Laulo nuoren Joukaisen Suohon suoni vöistä Niittyhyn nisu lihoista Kainalosta kangahasen Laulo koiran Joukahaisen Kynsin kylmäh kivehen Hampahin vesi hakoh Laulo jousen Joukahaisen Kaariksi vesien päälle Laulo nuolen Joukavaisen Haukaksi kiitäväksi [He] sang young Joukahainen Into a swamp to his belt Into a meadow to his waist Into a heath up to his armpit Sang Joukahainen’s dog’s Claws into a cold stone Teeth into a water-log Sang Joukahainen’s bow Into [to become] a (rain)bow on the water Sang Joukahainen’s arrow Into [to become] a streaking hawk 458 FROG Ylähäksi taivosella High in the heavens (SKVR I1 185.41-52) The series of elements in this multiform can be considered symbolic equivalents. They may be seen as distinct, but they are simultaneously elements of a coherent symbolic integer. Formally, semantic parallelism in the second part of each element exhibits a decreasing progression. In the first element, there are three parallel verses, followed by two in the second, one in the third, and one in the fourth. The end of the pattern is marked in the fourth element with an additional line that is not semantically parallel. The pattern is illustrated in the semantic diagraph in (23). Mention of the teeth // claws of the hero’s dog stands out in the diagraph, and should be regarded as a consequence of using a conventional couplet as a pre-fabricated unit: (23) into landscape location3 into landscape object2 hero’s bow to become rainbow hero’s arrow to become bird Sang hero Sang hero’s dog Sang Sang by teeth//claws in sky The progression in this variant is not so even in all performances in this dialect. Nevertheless, it remains apparent that the weight of duration is toward the beginning. Elsewhere in this volume, David Holm and Eila Stepanova observe that the amount of parallelism is an indicator of significance in the traditions they analyse. Here, the gradual decrease of parallelism seems to be a technique to keep emphasis on the first element, the fate of the hero. Subsequent elements reinforce the first element through symbolic parallelism without competing with its prominence. The hero’s dog, bow, and arrow are not otherwise mentioned in the epic, which underscores that their appearance serves to reinforce the symbolic integer of the hero’s dissolution into the landscape. 4.4. Motif Parallelism with Macro-Parallelism Adjacent and proximate motifs may be organized in parallel structures and such parallelism is clearly situated at the level of symbolic articulation. In kalevalaic poetry, however, multiforms that operate as macro-formulae mirror and reinforce parallelism between the symbolic integers they signify. The recurrence of text sequences having the scope of multiforms produces macro-parallelism (Urban 1986:26-29). In §5.2, the hero’s dialogue with his mother at the beginning of The Song of Lemminkäinen is discussed. In that dialogue, the hero repeatedly demands his armor (translated literally as “war-shirt”), until his mother complies. The command and compliance are expressed through the same multiform with morphological variation of pronouns, possessive affixes, verb inflections, and in the first couplet also particles. Elements that vary between these uses are shown in italic font in (24-25): (24) Oi emońi vanhempańi Tuo sie miun sotisomańi Oh my mother my elder Bring my war-shirt PARALLELISM DYNAMICS I (25) Kanna vainovoattieńi Pitoloissa piettäväńi Häissä häilyteltäväńi Carry my persecution-garb My holding-one at the feast My glistening one at the wedding Jo emonsa vanhempanša Jo tuopi hänen šot’isomansa Kanto vainovoattiša Pitoloissa piettävänšä Häissä häilyteltävänsä Already his mother his elder Brings indeed his war-shirt Carries his persecution-garb His holding-one at the feast His glistening one at the wedding (SKVR I2 791a.139-48) 459 Linguistic macro-parallelism in command-compliance interaction is certainly not unique to kalevalaic epic (for example, Urban 1991:71-72). In this case, it produces cohesion between the command and compliance and performs a rhetorical function. It marks the conclusion of a longer dialogue in which this multiform is expressed six times (on its variation in reproduction, see Frog 2016a:90-91). Combined motif parallelism and linguistic macro-parallelism may also be observed, for example, in the paired motifs of weeping associated with the hero’s departure from sexual adventures on the island of women. In the dialect of the example below, the hero is called Kaukomieli. The description of his grief in (26) is presented through a schema of negative parallelism with a) the introduction of the motif of the hero weeping for a period of time, b) a negative statement about its cause, and c) a positive solution. The motif has a strong visual component. The hero does not simply weep: he weeps as long as he can see the soaren puut (“trees of the island”), and the trees are gradually identified as emblematic of the women living on the island: (26) Sini itki Kaukomieli Kuni soarem puut näküvi Ei se itke soarem puita Eikä itke soarem maita Itki soaren impilöjä So long wept Kaukomieli As long as the island’s trees could be seen He did not weep for the island’s trees Nor wept for the island’s lands Wept for the island’s dear maidens (SKVR I2 791a.292-96) This is paralleled by a corresponding motif of women on the island weeping for as long as they can see the masts of the hero’s departing ship in (27). The correspondence of motifs is matched in verbalization: both motifs are built from the same linguistic template of structure and phraseology. The Karelian word for “mast,” purjehpuu, literally translates as “sail-tree,” producing a lexical correspondence that correlates the different types of “trees” in each motif. (27) Sini itki soaren immet Kuni purjeh-puut näküvi Rauta-hankat haimenta₍a Ei ne itke purjeh-puita So long wept the island’s maidens As long as the masts [lit. “sail-trees”] could be seen The iron yard dimmed They did not weep for the masts 460 FROG Rauta-hankoja halaja Itki purjeh-puullisia Rauta-hankan haltieta Wish for the iron yard Wept for the mast-people [sailors] The iron yard’s tenders (SKVR I2 802.297-303) The sort of shared verbal framework observable between (26) and (27) is quite familiar in kalevalaic epic. It reflects a common multiform with open slots that are completed in different ways, illustrated in (28): (28) Sini itki X Kuni Y puut näküvi [PARALLEL VERSE] Ei [PRONOUN] itke Y puita [PARALLEL VERSE] Itki Y Z [PARALLEL VERSE] So long wept X As long as the Y trees could be seen [PARALLEL VERSE] [PRONOUN] did not weep for the Y trees [PARALLEL VERSE] Wept for the Y Z [PARALLEL VERSE] X = a four-syllable name or expression for the grammatical subject Y = a two-syllable noun with a long stressed syllable as the qualifier for “trees” Z = a four-syllable word that identifies people in the location of the “trees” Although both descriptions are analyzable as variations on an open-slot multiform, the open-slot multiform is only generative in principle. Hypothetically, it could be used to identify the abstract motif with other agents in different situations,56 but, to my knowledge, there is no evidence of this. In practice, the open-slot multiform was a framework shared across two motifs, but the different ways that it was completed crystallized into independent multiforms. In other words, the slots were not completed anew in each situation: they crystallized into series of verses from which potential parallel verses may be omitted. Rather than being generative, the shared framework’s significance is the production of macro-parallelism between (26) and (27), represented in the semantic diagraph in (29): (29) Kaukomieli weep Maidens weep duration: sees island cause ≠ island2 duration: see ship2 cause ≠ ship2 cause = maidens cause = sailors2 (i.e. Kauko) Macro-parallelism creates cohesion across the two motifs while highlighting variations between them. These may yet vary in length (in reproduction) through the presence or omission 56 For example, it could be hypothetically used in a conclusion to The Courtship Competition, an epic in which two hero’s compete for the maiden of Pohjola and one of them loses. Such an adaptation would only require completing X with the losing hero’s four-syllable name (Väinämöinen or Ilmarinen), Y with Pohjo (genitive Pohjon) as a common two-syllable variation on Pohjola, and the plural impilöjä could be used in the singular with a diminutive ending, impilöini, partitive impilöistä. It is also quite possible that the multiform was used outside of epic narration with reference to, for example, a living person (X), completing Y and Z to refer to an actual or potential romantic situation—but this was not the sort of poetry that collectors were interested in documenting. PARALLELISM DYNAMICS I 461 of parallel verses, but the multiforms construct equivalent metered frames for their respective symbolic integers. The macro-parallelism also schematically structures the metered frames to produce formal parallelism in symbolic articulation. This structuring advances parallels from constitutive elements of the respective motifs to their relationships. Correspondences between individual features and variation are brought into focus within those structured relations. Parallelism is developed at multiple orders of signification simultaneously. Linguistic macroparallelism makes motif parallelism more salient. If focus is exclusively on linguistic macroparallelism, (26) and (27) may be viewed as a variety of semantic parallelism of the sadness of the scene. When focus is on symbolic articulation, symbolic parallelism exhibits a network of correlations and contrasts that do not reduce to semantic or analogous equivalence per se. 4.5. Motif Parallelism without Macro-Parallelism Motif parallelism is not dependent on systematic linguistic macro-parallelism. In kalevalaic epic, “recurrent returns” of motifs tend to be made salient through lexical and structural recurrence. The sort of highly structured parallelism in symbolic articulation illustrated in the preceding section is matched with parallelism in verbalization. Manifesting comprehensive macro-parallelism in such cases is simply part of how the tradition works. Stepping away from the kalevalaic epic tradition for a moment, a stanza from the end of a medieval Scandinavian ballad Liten Karen (“Little Karen”) provides an illustration of how parallelism in symbolic articulation can diverge from verbalization. In this ballad, a virtuous maiden has died at the hands of a wicked king or duke. The following stanza describes the ultimate fate of each character, that of the maiden in the first couplet and of the duke in the second: (30) Der kommo tvanne dufvor fran himmelen neder Men nar de flogo dadan de syntes vara tre Der kommo tvanne korpar fran helvetet upp De togo unga Hertingen bade med sjal och kropp There came two doves down from heaven But when they flew thence they seemed to be three There came two ravens up from hell They took the young Duke, both body and soul (Text and trans. from Sands 2001:349) In this case, line 3 exhibits lexical reproduction from line 1 with grammatical parallelism. Words that are not reproduced from line 1 form contrasts between the types of bird dufvor // korpar (“doves // ravens”), the locations himmelen // helvetet (“Heaven // Hell”) from which they come, and the deictics neder // upp (“down // up”) of their directions of movement. These contrastive pairs are powerfully charged with morally encoded cultural symbolism. The linguistic parallelism of these verses is shown in the diagraph in (31): 462 FROG (31) Der kommo tvannne dufvor fran himmelen neder Der kommo tvannne korpar fran helvetet upp There came twoo doves from Heaven down There came twoo ravens from Hell up Like in the example of the hero and the maidens weeping (26-27), linguistic units produce a metered frame for elements in symbolic articulation, and linguistic parallelism structures symbolic elements within that frame making their relationships salient. The verses establish the doves as agents of Heaven and the ravens as agents of Hell. Motif parallelism continues and is completed in lines 2 and 4. Though these lines lack linguistic parallelism, the salience of motif parallelism in lines 1 and 3 establishes parallelism as a frame of interpretation for the second member of each couplet. The contrastive parallelism of lines 2 and 4 is embedded in cultural knowledge about the symbolism being manipulated. The interpretation of three rather than two birds returning to Heaven relies on recognizing the bird as a potential image of the soul in conjunction with conceptions of what happens to the soul following death. The contrast between the apparently willing flight of the doves and the aggressive “taking” of the Duke “body and soul” also engages cultural conceptions. It relies on the mythic image of Heaven as a desirable location and that of Hell as a place of punishment and suffering to which people do not go willingly. On that background, a diagraph of elements in symbolic articulation makes the parallelism visible: (32) Line 1 Came 2 doves down from Line 3 Came 2 ravens up Line 2 Flew (3rd = Line 4 Carried from Heaven Hell maiden) back (to) (Heaven) = grace the Duke (+ ravens) (back to) (Hell) = punishment three Symbolic articulation is here organized in the metered frames of lines and couplets and characterized by the juxtaposition of image elements and motif constituents. The juxtaposition informs the significance of each element through its identification or contrast with a counterpart in the parallel motif. Within the tradition, the motifs are placed on an axis between Heaven and Hell in a contrast of cosmological proportions. As a more complex unit of narration, motif parallelism can manifest in the juxtaposition of multiple elements at the level of symbolic articulation. Formal aspects of verbalization, such as the reproduction of lexical items from the preceding utterance, grammatical structures or formal units such as lines and couplets, may make the parallel structure more salient. Nevertheless, the parallelism manifested is at the level of symbolic articulation, reflecting a unit of narrative content. Motif parallelism may, potentially, occur in kalevalaic epic without linguistic macro-parallelism, but this involves questions of how parallelism should be defined. When addressing motif-based parallelism within a multiform, examples of the hero’s response to a danger were built on elements of the hero’s action and the monster’s counter-action, which indicated its success. Similarly, command-compliance interaction clearly manifests as PARALLELISM DYNAMICS I 463 parallelism when it is matched by linguistic macro-parallelism. The question then arises whether the image of the danger on the hero’s journey and his response manifest a parallelism. Multiform structures situate them in equivalent metered frames. Lexical recurrence was noted above and is connected with a recurrent return to the image of the eagle at the conclusion of the motif. These features create cohesion between the two narrative integers: the image of the eagle precedes and concludes the hero’s action, framing it. The symbolic integers, paired as a threat and response, are also matched and made coherent through the organization of metered frames and recurrence in verbalization. Symbolic parallelism between motifs is often difficult to reduce to pure “semantic” parallelism. Parallel symbols do not necessarily mediate “the same” symbolic or semantic content; instead, they construct and develop meaningfulness and significance through the relations of symbolic elements. As a consequence, describing motif parallelism as “semantic parallelism” or even “analogous parallelism” may be misleading. The relations between examples (26-27) would be better described as “corollary symbolic parallelism,” and that between motifs in (30) as “contrastive symbolic parallelism.” Correlation and contrast are essential to both types, the difference is what comes into focus. These types of parallelism are comparable to linguistic semantic and contrastive parallelism operating at the level of symbolic articulation. Parallelism between the eagle and the hero’s response to it, on the other hand, creates a rhetorical connection and strong cohesion between complementary integers. If an analogue is made to verse parallelism, rather than a semantic base, this type of motif parallelism is more comparable to forming parallel groups with recurrent elements in an additive or complementary series, such as the description of the eagle in its tree or the alder-wood man and his boat. Motif parallelism may, it seems, be able to take as wide a variety of forms as does verse parallelism. 5. Higher Order Parallelism Higher order parallelism occurs when images and motifs communicate complex symbols that are socially recognizable within a tradition community, and when those integers of tradition mediate still higher order signs.57 Discussion here begins with parallelism at the level of themes, looking at themes that are reproduced in series within a more complex episode. Example (33), used for this discussion, provides the basis for looking at parallelism between that episode and another within the same epic. The emphasis on formal aspects of higher order parallelism in these examples will advance to encompass the construction of the meaningfulness or significance of symbols participating in higher order parallelism. 57 The approach to higher order parallelism outlined here has developed especially through discussions with Karina Lukin, who addressed parallelism in larger units of text in Nenets epic at the seminar that gave rise to the present volume (Lukin 2014:122-29). 464 FROG 5.1. Theme Parallelism A theme was defined above as an integer of tradition that is mediated through conventionally associated images, motifs and/or equivalent sets of these. In kalevalaic epic, theme parallelism within an episode is particularly noticeable. The recurrent images and motifs comprising the theme are expressed through the same multiforms, producing macro-parallelisms. Sequential recurrence of a theme with variation resembles recurrence of an open-slot multiform at a higher level of structure. Its variable slots can be embedded in recurrent open-slot multiforms, slots for whole multiforms, or both. Because themes in this tradition are inclined to form a structured and organized series of images and motifs that each forms a metered frame, they can operate at the level of symbolic articulation, much as a macro-formula does at the level of verbalization. Example (33) is taken from the hero’s journey in The Song of Lemminkäinen. The first of the three threatening “deaths,” the fiery eagle, is followed by a giant “worm” or serpent, and “wolves//bears” in iron “bridles//shackles.” Each danger is presented with the hero’s response to it in one recurrence of the theme template organized within the narrative pattern. A formulaic “boundary marker” (Lamb 2015:236) opens the narrative pattern at the beginning of the first theme: läksi Päivöilän pitoho / hyvän joukon juominkihi (“Left for the feast of Päivölä / The good group’s drinking-feast”). The grammatical subject is omitted from the first use of the theme’s opening multiform; the naming formula Lemminkäini lieto poika (“Lemminkäinen lieto boy”) is used as a boundary marker at the beginning the theme’s subsequent iterations. Following the opening multiform, the theme template has an open slot for the danger faced by the hero, and then for the hero’s response. The first and third instances express the hero’s response with a common open-slot multiform; all uses conclude with the same couplet. The third iteration is followed by repeating the boundary marker couplet with variation in the verb, completing the episode: peäsi Päivölän pitoho / hyvän joukon juominkihi (“Got to the feast of Päivölä / the good group’s drinking-feast”): (33) The example of the hero’s dangerous journey from The Song of Lemminkäinen is here laid out with each recurrence of the theme in one of three columns. Texts have been arranged in a manner similar to a diagraph with parallel multiforms and their textual features aligned. Danger 1 Danger 2 Danger 3 Lemmingäińe liedo poiga Mäńi matkoja vähäizen Käüt on teidä pikkaraizen N’iin on kuin sano emoińe Oma vanhembi vakitti Lemmingäińe liedo poiga Mäńi matkoja vähäizen Käüt oñ teidä pikkaraizen N’iin on kuin sano emoińe Oma vanhembi vakitti Läksi Päivöil’äm pidoho Hüvän joukon juomingihi Mäńi matkoja vähäizen Kävi teidä pikkaraizen Niin on kuin sano emoińe Oma vanhembi vakitti PARALLELISM DYNAMICS I Danger 1 Danger 2 Tuliba tulińe koski Kozess’ oñ tulińe luodo Luuvoss’ oñ tulińe koivu Koivuss’ oñ tulińe kokko Üöt heän küñzieh hiveli Päivät kaikki kitskutteli Lemmingäizem peäñ varalla Mad’ on t’iellä poikki-puolin Küllöil’än kujojen suissa Pitelämb’ om pert’in hirttä Suzit oñ rauda-suittsiloissa Paksumb’ om pert’im patsasta Karhut rauda-kahlehissa Lemmingäizem peän varalla “Kokkoizeńi, linduizeńi “Mado musta moan alaińe Toukka tuomen karvallińe Luo sie silmäs luodehella Keännä peäzi päivän alla Laulan karjan teürilöidä Süüväksezi kokkoizeńi Piśśä peäzi mättähähä Itse kuvote kuloho Ne on sü
öt so
am miestä Tuhonnut tuhat urosta Kokkoizeńi, linduizeńi Anna männä matka-miehen Lemmingäizeñ liijatengi” Danger 3 Anna männä matka-miehen Lemmingäizen liijatengi” “Ohtoizeńi, linduizeńi Ohtoińe metsän omena Mezi-kämmen källeröińe Luo sie silmäs luodehella Keännä peäzi päivän alla Laulan karjañ lambahie Süüväksezi ohtoizeńi Ohtoizeni, linduizeńi Ohtoińe metsän omena Mezi-kämmen källeröińe Anna männä matka-miehen Lemmingäizen liijatengi” Peäzi Päivoil’äm pidoho Hüvän joukon juomingihi (SKVR I2 717.75-95.) (SKVR I2 717.96-109.) (SKVR I2 717.110-34.) Lemminkäinen lieto boy Went the journey a little Walked is the road a bit Thus is as dear mother said [His] own parent advized Lemminkäinen lieto boy Went the journey a little Walked is the road a bit Thus is as dear mother said [His] own parent advized Left for the feast of Päivölä, The good group’s drinking-feast. Went the journey a little Walked the road a bit Thus is as dear mother said [His] own parent advized Came indeed a fiery rapids A worm is across the road In the rapids is a fiery islet Longer than a house’s beam In the islet is a fiery birch Thicker than a house’s post In the birch is a fiery eagle By night hones his claws Every day strutting slowly Ready for Lemminkäinen’s head At the ends of Kyllölä’s paths Wolves are in iron bridles Bears in iron shackles Ready for Lemminkäinen’s head They have eaten a hundred men Destroyed a thousand heroes 465 466 FROG Danger 1 Danger 2 Danger 3 “My dear eagle, my little bird “Black worm, under the earth “My dear bear, my dear bird Death’s furry larva Dear bear, fruit of the forest Honeypaw, brown one Cast your eye to the northwest, Stick your head in a tussock Cast your eye to the northwest Turn your head beneath the sun Knit yourself into a brushfire Turn your head beneath the sun I sing a flock of black grouse I sing a flock of sheep For your eating my dear eagle For your eating my dear bear My dear eagle, my little bird Let go the travelling man Lemminkäinen oh indeed” Let go the travelling man Lemminkäinen oh indeed” My dear bear, my little bird Dear bear, fruit of the forest Honeypaw, brown one Let go the travelling man Lemminkäinen oh indeed” Got to the feast of Päivölä The good group’s drinking-feast Each of the three dangers is highly conventional. The second danger presents only the first element of the relevant multiform. Both elements of this multiform are used in the preceding dialogue between the hero and his mother. There, the second element is “Se on sü₍öt so₍am miestä / Tuhonnut tuhat urosta” (SKVR I2 717.43-44) (“It has eaten a hundred men / Destroyed a thousand heroes”), like the second element in the third danger’s multiform, and followed by “Lemmingäizem peän varalla” (717.45) (“Ready for Lemminkäinen’s head”). When this multiform was reproduced in the narration of the journey, it was truncated. Shortening may have continued through the completion of the theme. The hero’s response to the danger is consistently structured across the three encounters: a vocative address to the danger, followed by verbalization of a magical act for overcoming the danger. In other dialects of the tradition, the closing couplet can appear to be an integrated part of a response multiform. In this case, its status is less clear. My initial inclination was to view the reproduction of the vocative addresses to the eagle and the wolves//bears as a continuation of the preceding line in extended parallelism (Frog 2014b:197-98). However, vocative formulae tend in general to open a new unit of utterance in the epic register. If this is the case here, the vocative formula is an indicator that the singer treats the final couplet as a discrete unit, a closing boundary marker for the theme. This unit then has the potential of a multiform to either be expanded with a vocative address or to appear alone, as in the response to the worm. Whether this boundary marker should be treated as a discrete multiform remains ambiguous. As with open slot multiforms, the potential of the theme as a generative template is not customarily realized. However, Viena Karelia, where the variant in (33) was recorded, was unusual in that migrations in preceding centuries had brought several dialects of the tradition into the same region. Some singers capitalized on knowledge of dialectal variants to incorporate more than three dangers on the journey. 58 The incorporation of additional dangers integrates them into the dialectal form of the theme. Such innovations would involve use of the theme as a generative 58 A chart of dangers encountered by the hero covering most variants can be found in Frog (2010a:389-95). PARALLELISM DYNAMICS I 467 template, whether generations earlier or in the situation of performance (for example as a display of knowledge and skill for a collector). Each use of the theme has a simple three-part structure: a) a multiform indicates progress of the journey until the encounter; b) a multiform represents a danger; c) a multiform or pair of multiforms present the hero’s response to that danger. The following diagraph illustrates this three-part structure as well as the three types of boundary markers as found in (33):
 Departure Encounter Danger Response Boundary marker 1 Travel Eagle Sing A Boundary marker 2 Travel Worm Curse Boundary marker 2 Travel Wolves/bears Sing B Arrival Boundary marker 3 The recurrent reproduction of the same multiform at the beginning of the theme and the same couplet at its conclusion combine with structural recurrence to align the metered frames in parallelism. The mythic images of the three dangers and the responses to them are aligned as symbolically parallel integers within the theme. This motif parallelism is reinforced by linguistic macro-parallelism in the hero’s responses to the first and last dangers. Thematic parallelism imbues each varying element of the theme with equivalent significance in the same way that nails//teeth exhibit equivalence across parallel verses in (4) above. Theme parallelism is quite common in kalevalaic epic dialogues. A sequence of dialogue is often organized as a series of turns forming a theme. The theme is organized around a request or demand and a response between two interlocutors. Some cycles of dialogue recur with variations until a particular variation, such as compliance with a demand, allows the narration to proceed. Dialogic theme parallelism is an organizing principle of the hero’s dialogue with his mother at the beginning of the epic, when she warns him of the “deaths” on the road. The most common structure is: a) the hero demands his armor (or some equivalent thing); b) his mother forbids him from going to the feast and warns him of a danger; c) the hero rejects the threat’s validity, perhaps explaining how he will overcome it; and then a) the demand is repeated. In practice, the theme may be concentrated into three turns or expanded into several. For example, the hero can ask what the danger is after his mother forbids him from going, breaking one of her turns into two parts. In the example quoted above, the structure is slightly different. The hero simply begins preparations in a boundary-marker multiform (BM 1) that begins the episode. His mother makes a prohibition, stating that there are three “deaths,” and continues directly into a description of the first danger. The hero dismisses her warning and asks her what is next. In the second and third uses of the theme, another multiform is spoken by the mother as an opening boundary marker (BM 2). Following the rejection of the third danger, the hero demands his armor and the mother complies, concluding the episode with the command-compliance parallelism illustrated in (24-25). This structure of this example is illustrated in (35): 468 FROG (35) Hero’s Decision Mother’s Prohibition Warns of Danger Hero’s Response BM 1 3 “deaths” Eagle Reject BM 2 Worm Reject BM 2 Wolves//bears Reject Hero’s Demand Mother’s Response Shirt Shirt The salience of theme parallelism in kalevalaic epic is exceptional while theme parallelism itself is common in many narrative traditions. In other traditions, this type of higher order parallelism may be less transparent at the level of verbalization and its metered frames may be more flexible. 5.2. Episode Parallelism “Episode” is here used as a practical term for the unit of narration above the level of a theme in this tradition. As discussed above in §§1.2-3, an episode is characterized by a narrative pattern as a higher order integer, which in kalevalaic epic is mediated through conventionally associated themes. Parallelism can also occur between episodes. Episode parallelism is less common than theme parallelism and seems to manifest more frequently at the level of structures with different themes rather than reproducing the same themes in different episodes. The hero’s dialogue with his mother about dangers on his journey and narration of the journey itself exemplifies this point. The episodes of the dialogue and the journey exhibit structural parallelism in the triple recurrence of a theme as well as the specific arrangement of boundary markers. In both episodes, the three dangers are iterated through consistent multiforms in the same sequential progression across the three recurrent themes. In addition, the second boundary marker is the same multiform as the opening description of travel, but morphologically varied to be spoken by the hero’s mother. As is common in this epic, the recurrent structures and patterns in the symbolic integers organize metered frames that make the two episodes perceivable as parallel members of a group, illustrated in (36): (36) Instigation Theme 1 Theme 2 Theme 3 Conclusion Preparation Danger 1 Danger 2 Danger 3 Preparation completed Departure Danger 1 Danger 2 Danger 3 Arrival The otherwise unusual recurrence of both the multiform of the danger and the multiform of travel in this variant enhances the salience of parallelism by opening themes in both episodes with the same multiforms in series. When both episodes are included in a performance, the reproduction of multiforms across episodes may be limited to those expressing the images of the dangers. In their context, it is not clear that their recurrence across episodes was perceived as linguistic macro-parallelism per se. It would certainly produce resonance and cohesion between the episodes, linking these symbolic elements in series. However, just as not every use of a formula is perceived as a parallelism, the recurrent multiform may have been perceived PARALLELISM DYNAMICS I 469 practically as how the eagle must be described when mentioned again. Nevertheless, the episode parallelism produces a strong cohesion, reinforced by the fact that the journey invariably follows immediately on the dialogue. The dangers of the journey were the core of the epic, although they could be presented exclusively in the dialogue, with a jump in narration from the hero’s departure to his arrival at the feast. Nevertheless, the journey was not presented without being preceded by the dialogue, unless it was separated from the epic context and performed as part of an incantation.59 The fact that the paired episodes could be reduced to a single episode with priority on the dialogue emphasizes the cohesion between them. Episode parallelism could also occur between the hero’s dialogue with his mother and another later dialogue in the epic. If, rather than being killed, the hero slays the host of the feast of Päivölä in a duel, he flees from vengeance and ends up going to the island of women. In this case, however, the hero does not flee directly to the island. Instead, he first returns home and asks his mother where he should hide. In the ensuing dialogue, his mother proposes a series of locations, each of which he rejects until she recommends the island. This is the second episode organized as a dialogue between the hero and his mother in the epic, and both dialogues are similarly structured. Both customarily open with the hero addressing his mother with a request. The core of the dialogue is the mother giving advice and the hero rejecting it. Complexity of the recurrent themes in these dialogues and some aspects of their organization vary across dialects. The example above is a case in point: the hero’s request for his armor generally opens a theme of that dialogue rather than only concluding it. Nevertheless, the core remains stable as does the structuring of the episode around dialogic theme parallelism. Both episodes structurally occupy equivalent positions in a larger narrative pattern. Each episode follows an instigating event at the beginning of a heroic adventure as a larger structural unit within the epic. Within that larger unit, the dialogue immediately precedes the hero’s departure to the otherworld location of that adventure. The diagraph in (37) illustrates structural and symbolic correlations of the two episodes that produce an episode parallelism:60 (37) Instigation Dialogue Theme 1 Theme 2 Theme 3 Conclusion Hero seeks mother with Mother gives request (armor) advice Reject Reject Reject Hero departs on journey Hero seeks mother with Mother gives request (advice) advice Reject Reject Accept Hero departs on journey The two dialogues differ in two relevant respects. In the first, the mother asserts her advice in the form of a prohibition and warning; in the second, the hero seeks out her advice. The 59 On The Song of Lemminkäinen in or as an incantation, see Frog (2010a:75n.96, 79-80, 82, 84, 86-87), and see also Frog, “Parallelism Dynamics II,” this volume. There are a very few examples that could reflect performances without the dialogue, but they are only short fragments of a few lines and remain either ambiguous or generally unusual. 60 The variation in the example in (36) would alter the unit under “Instigation” from “Hero seeks mother with request (armor)” to “Hero prepares for departure.” In this case, the correspondence would shift to “Hero prepares to flee” in the second episode. However, the performer of that variant did not continue the epic to include the duel and escape to the island, so the variation there had no relevance for episode parallelism. 470 FROG second difference is structurally dependent on the first, because it makes them dialogues of subtly different types: in both, a variation occurs that allows the plot to move forward, but it occurs at different points and thus gives the dialogue a different structure. In the first, the recurrent theme’s structure is uniform and the variation occurs at the conclusion of the third cycle, when the mother complies with the hero’s request to fetch his armor. The mother’s compliance is often preceded by a reiteration of the hero’s request that could otherwise begin a recurrence of the dialogue sequence. However, the result is the command-compliance parallelism as in (24-25) above. The multiform produces resonance with its previous uses, when the demand is incorporated into the dialogue. Nevertheless, the command-compliance parallelism of the same multiform used in series is a use of macro-parallelism that contrasts with the dialogic theme in structure, scope, and significance. Theme parallelism does not manifest, and the variation instead marks the conclusion of the episode. The second dialogue is structured so that the variation must occur within the theme with an acceptable response. This response may lead into an alternative conclusion to the theme, or the theme may conclude abruptly. In the latter case, the initial turns of the dialogic theme are sufficient to activate perception of the recurrent theme’s metered frame, even though it lacks an integer in the place of the hero’s rejection of the proposal. The effect is a form of catalexis, when a verse is truncated so that not all metrical positions are completed. The parallel members of this episode’s “recurrent return” may be separated by a considerable amount of narration (at least for a short epic form), often 50-100 verses. This amount could be doubled in a performance where every verse is sung twice. The scope of the episodes facilitates the salience of parallelism in a way that smaller units would not. Reproduction of a single line or couplet following such an interlude would not necessarily be noticed at all. A corresponding recurrence of a single motif might produce resonance without being perceived as parallelism in the on-going flow of discourse. The dialogic theme parallelism within each episode foregrounds form and, thereby, resonance so that parallelism between the episodes is more easily perceived. However this example of episode parallelism is interpreted, it serves to link the adventure sequences of the epic and reinforce cohesion between episodes. 5.3. Parallelism and Meaning Construction Higher order parallelism can be developed to structure the meaningfulness of images and motifs within parallel members. The so-called Sampo-Cycle of kalevalaic epic exhibits a range of variation, but can be generally described as (or historically as having been) an account of the creation and organization of the world.61 The epic centers on the creation of a mysterious prosperity-creating object, a sampo (interpreted in diverse ways62), its theft from the otherworld, and destruction as the heroes seek to escape with it. Parallelisms appear between the creation of 61 The forms of this epic have great variation by region. The fundamental studies remain Setälä (1932:169-607); Harva (1943); Kuusi (1949); in English, see also Kuusi (1994) and Frog (2012a:222-40). 62 See Setälä (1932:169-91) and Kuusi (1949:142-48); in English, concentrating on Viena Karelia, see Tarkka (2012:passim). PARALLELISM DYNAMICS I 471 the world from an egg at the beginning of this epic cycle and the events surrounding the destruction of the sampo in a conflict with a bird-formed adversary. In the world-creation, a bird flies about, looking for a place to build a nest. The demiurge Väinämöinen raises part of his body as the first land, where the bird builds a nest and lays eggs. Owing to the uncomfortable heat of the bird’s brooding, Väinämöinen moves and the eggs roll off (and the bird flies away). One egg breaks, and from it Väinämöinen creates the world, or parts of it (especially the celestial bodies). At the conclusion of the cycle, a group of heroes led by Väinämöinen steals the sampo from the otherworld and escapes by sea. The Mistress of Pohjola (“North-LOCATION”) pursues in her own ship. Väinämöinen intervenes by throwing back a piece of flint, magically causing a shoal or skerry to rise from the sea behind them. The pursuing ship runs aground and breaks apart. The Mistress of Pohjola transforms into a magical bird, which may be explicitly formed from her ship and crew, and continues her pursuit in flight. She lands on Väinämöinen’s ship and takes hold of the sampo, intending to flee with it. Väinämöinen strikes, the sampo breaks, and the adversary flies back to the north (with part of the broken sampo). Prosperity and fertility are distributed in the world with the sampo’s fragments, establishing the principles of so-called “limited good” (Tarkka 2012:154-59). The parallelism between the creation of the world and the destruction of the sampo occurs at the level of themes, each of which is embedded in a more complex narrative pattern. Although this summary or even a transcription of an epic performance may not make the parallelism transparent and evident, it surfaces when presented as a diagraph:63 (38) Bird flies Lands on V.’s knee Egg on knee (bird lays) Adversary flies as bird Lands on V.’s ship Sampo on ship (bird takes) V. World-egg moves breaks V. strikes Sampo breaks (Bird departs) World is created Adversary departs Prosperity distributed Higher order parallelism correlates symbolic elements that are aligned within metered frames. Here, correlation with the mythic egg of creation image informs the significance of the sampo by generating a symbolic equivalence between them. This symbolic equivalence also manifests in the interpretations of singers, as witnessed by Ohvo Homanen’s explanation that: “Sammossa oli kaikki moailman hüvüs, kuu ja päivä ja armas aurinkoinen” (SKVR I1 83a, n.19) (“All of the goodness in the world was in the sampo, the moon and daylight and dear sun”) (see also Tarkka 2012:149). Whether understood literally or symbolically, the statement indicates a correlation between the sampo and the world-egg, from which the sun, moon, and celestial sphere are created in the epic. Symbolic correlation is marked by contrasts. Whereas the bird of creation is a positive agent, the bird in the battle over the sampo is the Mistress of Pohjola, a powerful adversary of the heroes. Väinämöinen seeks to help the bird of creation but must defend the sampo from the Mistress of Pohjola. The breaking of the world-egg is a positive event that creates the cosmos; the breaking of the sampo is a tragedy. Its destruction distributes prosperity, much of which is 63 I have previously discussed the referential relationships between these parts of the Sampo-Cycle without using the concept of parallelism or diagraph analysis in Frog (2012a:229). 472 FROG lost into the sea and a significant part of which is carried to the otherworld. Only fragments of the sampo make it to the world of the singers, otherwise crops would grow without sowing and there would, presumably, always be enough for everyone (see Tarkka 2012:155-57).64 These contrasts construct the bird-formed mistress of the otherworld as an antithesis of the bird of creation. Rather than introducing the egg as a mythic object with boundless creative potential, she seeks to remove its equivalent from the world. Parallelism between these narrative sequences informs the significance of the battle, expanding it to cosmological proportions. The mythic images and motifs of the world-creation and the battle over the sampo remain distinct. Correspondences at the level of verbalization sometimes produce resonance between elements in the parallel themes. When such resonance occurs, it may complement parallelism between symbols, such as the bird. However, the tradition does not exhibit a tendency toward linguistic macro-parallelism across the themes that would make the symbolic parallelism salient. There is no reason to assume that everyone was sensitive to this theme parallelism. The fact that the parallelism appears to have been maintained at a social level in regions where the parts of the cycle had not comprehensively transformed or broken down65 suggests that specialists and authorities in the transmission of kalevalaic epic recognized the relevant theme parallelism. When this theme parallelism is recognized, resonances with the preceding theme of the destruction of the otherworld ship also become perceivable as parallelism: (39) Bird flies V. raises knee/land Bird lands on knee World-egg breaks Boat pursues V. creates land Boat crashes into land Boat breaks V. creates world from egg (Bird departs) Adversary changes Bird pursues self/ship into bird Were it not for the parallelism in the battle over the sampo, correlations between the themes in (39) would likely remain resonant without being perceived as parallelism. Although there are correlations, the resonance extends outward from the emergence of land from the sea and something breaking against it. The motif of Väinämöinen raising earth from the sea to stop the pursuing ship employs a so-called “Object Flight” motif (Thompson 1955-58:D672): a protagonist or his magic helper throws something back that magically transforms to slow a pursuer. Elements of the description can also resonate with The Origin of Fire (Kuusi 1949:200). Both of these associations interfere with a clear and direct parallelism with the world-creation. The correlation is also not very strong between an egg breaking because it rolls from the nest and a boat breaking because it runs aground. The motif of creation through the transformation of something broken apart is an integrated part of kalevalaic mythology and does not itself produce a referential parallelism with The Song of Creation. The correlation of the bird of creation’s flight with the pursuit of the boat is contrastive and comes into focus only when a metered frame is 64 The contrast between an environment where a sampo produced all that was needed for subsistence and the reality of small villages with mixed-subsistence livelihoods was quite sharp in inland ecological environments near the Arctic Circle. 65 A significant transformation is that Väinämöinen as a demiurge is deleted from The Song of Creation in regions to the south (see Frog 2012a:222-26), which dissolves the parallelism discussed here. PARALLELISM DYNAMICS I 473 perceived. The correlation of the emerging bird that pursues the ship with the bird of creation relies on the stronger parallelism in the following battle. The theme of the boat’s destruction primes the parallelism in the following episode through resonances with The Song of Creation, and the latter parallelism reciprocally brings parallelism with the boat’s destruction into focus. As it does, the parallelism informs individual motifs with significance and emphasizes their power. To some degree, the contrasts help to establish the mistress of the otherworld as an antithesis of the bird of creation, but its salience should not be exaggerated. What is noteworthy is that the two parallelisms with The Song of Creation are not generally coordinated with each other as parallelisms. The destruction of the boat centers on the motifs of land being raised from the sea and a transformation from the broken object. The battle places emphasis on the contrast between creating the egg and rescinding the sampo on the one hand and the destruction and distribution of the sampo on the other. Both engage the creation event, but emphasize different aspects of it. The destruction of the ship and its transformation produces a parallelism with Väinämöinen’s agency in creating the world from the broken egg. The destruction of the sampo does not concern creative agency but the cosmogonic consequences of the object’s destruction. The two themes are parallel in their parallel engagements with the same theme of The Song of Creation, but accentuate different elements of it in different ways: (40) Event C-role V. raises Creation Flies Land Pursuit Sails Land Battle Flies C-role C-role Creates Lands Egg Crashes Lands -(sampo) V. acts Object Agent Creates Moves Breaks (Departs) World1&2 *Throws Breaks Strikes Breaks Bird Pursues Bird2 Departs Lim. good1 In addition to each theme parallelism appearing independent, it also appears that they are not in a symmetrical relation to the theme of The Song of Creation. In other words, parallelism manifests in the episode of the theft and destruction of the sampo so that the parts discussed here become perceivable as parallel members of groups with the relevant part of The Song of Creation. This parallelism seems to deeply inform the significance of the images and motifs surrounding the sampo. However, performance of The Song of Creation seems not to elicit a corresponding perception of parallelism with events surrounding the sampo: the significance of images and motifs of the world-creation are thus not reciprocally informed by the sampo’s theft and destruction (or at least not to a comparable degree). The relationship between these episodes thus exhibits a hierarchy, in which The Song of Creation is dominant and not necessarily affected by the engagement in parallelism. When parallelism becomes perceivable, the “recurrent returns” at higher orders of signification shape the meanings of the implicated symbols. The shaping of meanings emerges because parallelism produces a syntactic relation (in Morris’ sense) between participating signs. This syntactic relation distinguishes recurrent returns of parallelism from independent recurrence within a text. Theme parallelism in the Sampo-Cycle manifests as theme parallelism in part because the system of images and motifs is not a commonplace. Themes may recur in epics in many traditions, but parallelism occurs when recurrence produces a perception that the themes 474 FROG are parts of parallel groups. A theme or narrative pattern is itself meaningful if it is an integer of the tradition and operates as a sign. That sign will be meaningful through what John Miles Foley (1991:6-8) calls “traditional referentiality.” Traditional referentiality equates to the experiencebased understanding of a sign and its patterns of use, but does not link the formula, theme, or whatever back to any single particular use as parallel members of a group. Parallelism can be a powerful tool for the construction of meanings in a tradition, but not all recurrence manifests parallelism. 6. Perspectives Parallelism has been approached here as a broad semiotic phenomenon rather than being something exclusive to language. As such, parallelism can manifest in relation to metered frames at any level in signification. Most discussions of parallelism in oral poetry tend to concentrate on adjacent or proximate lines and groups of lines of verse.66 In contrast, the emphasis here has been placed on linguistically mediated signs and the correlations and interrelations of parallelism at different orders of signification. This approach highlights that parallelism both operates at different orders of signification and also interacts across them. Moreover, each level in signification conditions the metered frames of the signs it mediates. When one order of signs mediates another, it also participates in organizing elements within the metered frames of that next order of signs, which is structurally significant since these higher order signs are perceived as parallel members of groups. Macro-parallelism, on the other hand, is linked to recurrent returns in the signification of integers or elements at the next order of signification. As such, linguistic macro-parallelism appears simultaneously as a by-product of parallelism in symbolic articulation while simultaneously making that parallelism salient. Recognizing parallelism as a phenomenon that can occur simultaneously at multiple levels in signification offers new perspectives and helps bring these processes into focus. Du Bois’ diagraph analysis (§1.4) proves to be a valuable tool for analysing parallelism, whether it is applied to the lexical surface of text, its semantic components, or integers of symbolic articulation and higher orders of representation. The workings of parallelism across orders of signification are particularly easy to observe in kalevalaic epic owing to particular features of the tradition, yet the principles outlined here are readily adaptable to other traditions that are open to greater degrees of variation. The preceding discussion has advocated viewing parallelism as a fundamental and pervasive phenomenon. However, “parallelism” is an etic term: it is a word that we define and construct as a tool for research and analysis. As such, it is also flexible. Barring error and inconsistency with empirical data, any definition will be “right” or “wrong” only in relation to another definition, yet a particular definition may be better suited for certain investigations rather than others. This flexibility allows parallelism to be calibrated to particular research questions and materials. For some investigations, it will be most efficient to develop narrower models of parallelism that account, say, for the specific features of verse parallelism in a particular traditional poetry as an essential framework for analysing its uses and variations in specific texts 66 See Notes 2-3. PARALLELISM DYNAMICS I 475 (see the contributions of Holm and Saarinen elsewhere in this volume). The approach presented here is set apart by situating linguistic parallelism in relation to parallelism at the level of its primary mode of expression (alliteration, rhyme, etc.) on the one hand, and to parallelism at different levels of signs mediated through language (images, themes, and so on) on the other. The present model is oriented to parallelism in the broadest sense. Consequently, it is compatible with more narrowly specified and tradition-dependent models, and the more such models it is placed in dialogue with, the more it can be refined. Verse parallelism has received scientific attention since the mid-1700s. This study has shown that parallelism can be observed in metered frames of layers of signs mediated through language, opening an area where further research remains to be done. University of Helsinki References Austerlitz 1958 Robert Austerlitz. Ob-Ugric Metrics. Folklore Fellows’ Communications, 174. Helsinki: Academia Scientiarum Fennica. Berezkin 2015 Yuri Berezkin. “Folklore and Mythology Catalogue: Its Lay-Out and Potential for Research.” RMN Newsletter, 10:58-70. Berlin 1985 Adele Berlin. The Dynamics of Biblical Parallelism. Bloomington: Indiana University Press. Blommaert 2006 Jan Blommaert. “Ethnopoetics as Functional Reconstruction.” Functions of Language, 13.2:229-49. Blommaert 2007 . “Sociolinguistics and Discourse Analysis: Orders of Indexicality and Polycentricity.” Journal of Multicultural Discourses, 2.2:115-30. Cureton 1992 Richard D. Cureton. Rhythmic Phrasing in English Verse. London: Longman. Drout 2011 Michael D. Drout. “Variation within Limits: An Evolutionary Approach to the Structure and Dynamics of the Multiform.” Oral Tradition, 26.2:447-74. Du Bois 2007 John W. Du Bois. “The Stance Triangle.” In Stancetaking in Discourse: Subjectivity, Evaluation, Interaction. Ed. by Robert Engelbretson. Amsterdam: John Benjamins. pp. 139-82. Du Bois 2014 . “Towards a Dialogic Syntax.” Cognitive Linguistics, 25.3:359-410. 476 FROG Fabb 2015 Nigel Fabb. What is Poetry? Language and Memory in the Poems of the World. Cambridge: Cambridge University Press. Foley 1988 John Miles Foley. The Theory of Oral Composition: History and Methodology. Bloomington: Indiana University Press. Foley 1990 . Traditional Oral Epic: The Odyssey, Beowulf, and the Serbo-Croation Return Song. Los Angeles: University of California Press. Foley 1991 . Immanent Art: From Structure to Meaning in Traditional Oral Epic. Bloomington: Indiana University Press. Foley 1995 . The Singer of Tales in Performance. Bloomington: Indiana University Press. Foley 1999 . Homer’s Traditional Art. University Park: Pennsylvania University Press. Foley and Ramey 2012 John Miles Foley and Peter Ramey. “Oral Theory and Medieval Studies.” In Medieval Oral Literature. Ed. by Karl Reichl. Berlin: de Gruyter. pp. 71-102. Fox 1977 James J. Fox. “Roman Jakobson and the Comparative Study of Parallelism.” In Roman Jakobson: Echoes of His Scholarship. Ed. by Daniel Armstrong and C. H. van Schooneveld. Lisse: Peter de Ridder Press. pp. 59−90. Fox 1988 , ed. To Speak in Pairs: Essays on the Ritual Languages of Eastern Indonesia. Cambridge: Cambridge University Press. Fox 2014 . Explorations in Semantic Parallelism. Canberra: Australian National University Press. Fox 2016 . Master Poets, Ritual Masters: The Art of Oral Composition among the Rotenese of Eastern Indonesia. Canberra: Australian National University Press. Frog 2009 Frog. “Speech-Acts in Skaldic Verse: Genre, Formula and Improvisation.” In Versatility in Versification. Ed. by Tonya Kim Dewey and Frog. New York: Peter Lang. pp. 223-46. Frog 2010a . Baldr and Lemminkäinen: Approaching the Evolution of Mythological Narrative through the Activating Power of Expression. UCL Eprints. London: University College London. PARALLELISM DYNAMICS I 477 Frog 2010b . “Multiformit kalevalamittaisessa epiikassa.” In Kalevalamittaisen runon tulkintoja. Ed. by Seppo Knuuttila, Ulla Piela, and Lotte Tarkka. Kalevalaseuran Vuosikirja, 89. Helsinki: Suomen Kirjallisuuden Seura. pp. 91-113. Frog 2011a . “Alvíssmál and Orality I: Formula, Alliteration and Categories of Mythic Being.” Arkiv för Nordisk Filologi, 126:17-71. Frog 2011b . “Multiforms and Meaning: Playing with Referentiality in Kalevalaic Epic”. In Laulu kulttuurisena kommunikaationa—Song and Singing as Cultural Communication—Песня как средство культурной коммуникации: Proceedings from the Runosong Academy Jubilee Seminar, 8-10 October 2010. Ed. by Pekka Huttu-Hiltunen, Janne Seppänen, Eila Stepanova, Frog, and Riikka Nevalainen. Runolaulu-Akatemian Julkaisuja, 16. Kuhmo: Juminkeko. pp. 49-63. Frog 2012a . “Confluence, Continuity and Change in the Evolution of Mythology: The Case of the Finno-Karelian Sampo-Cycle.” In Mythic Discourses: Studies in Uralic Traditions. Ed. by Frog, Anna-Leena Siikala, and Eila Stepanova. Studia Fennica Folkloristica, 20. Helsinki: Finnish Literature Society. pp. 205-54. Frog 2012b . “On the Case of Vambarljóð I: Comments on Formulaicity in the sagnakvæði;” “II: Register and Mode from Skaldic Verse to sagnakvæði.” RMN Newsletter, 5:22-38, 49-61. Frog 2014a . “From Mythology to Identity and Imaginal Experience: An Exploratory Approach to the Symbolic Matrix in Viking Age Åland.” In The Viking Age in Åland: Insights into Identity and Remnants of Culture. Ed. by Joonas Ahola, Frog, and Jenni Lucenius. Helsinki: Academia Scientiarum Fennica. pp. 349-414. Frog 2014b . “Parallelism, Mode, Medium and Orders of Representation.” In Parallelism in Verbal Art and Performance: Pre-Print Papers of the SeminarWorkshop, 26th-27th May 2014. Ed. by Frog. Folkloristiikan Toimite, 21. Helsinki: Folklore Studies, University of Helsinki. pp. 185-207. Frog 2015a . “Mythology in Cultural Practice: A Methodological Framework for Historical Analysis.” RMN Newsletter, 10:33-57. Frog 2015b . “Registers of Oral Poetry.” In Registers of Communication. Ed. by Asif Agha and Frog. Studia Fennica Linguistica, 18. Helsinki: Finnish Literature Society. pp. 77-104. 478 FROG Frog 2016a . “Linguistic Multiforms: Advancing Oral-Formulaic Theory.” FF Network, 48:6-11, 14. Frog 2016b . “Linguistic Multiforms in Kalevalaic Epic: Toward a Typology.” RMN Newsletter, 11:61-98. Frog 2016c . “Metrical Entanglement and dróttkvætt Composition—A Pilot Study on Battle-Kennings.” In Approaches to Nordic and Germanic Poetry. Ed. by Kristján Árnason, Stephen Carey, Tonya Kim Dewey, Haukur Þorgeirsson, Ragnar Ingi Aðalsteinsson, and Þórhallur Eyþórsson. Reykjavík: University of Iceland Press. pp. 149-229. Frog and Stepanova 2011 Frog and Eila Stepanova. “Alliteration in (Balto-) Finnic Languages.” In Alliteration in Culture. Ed. by Jonathan Roper. Houndmills: Palgrave MacMillan. pp. 195-218. Halliday 1978 M. A. K. Halliday Language as Social Semiotic. London: Edward Arnold. Harva 1943 Uno Harva. Sammon ryöstö. Porvoo: Werner Söderström. Harvilahti 1992a Lauri Harvilahti. Kertovan runon keinot: Inkeriläisen runoepiikan tuottamisesta. Helsinki: Suomalaisen Kirjallisuuden Seura. Harvilahti 1992b . “The Production of Finnish Epic Poetry.” Oral Tradition, 7:87-101. Hasan 1989 Ruqaiya Hasan. Linguistics, Language, and Verbal Art. Oxford: Oxford University Press. Hasan 2007 . “Private Pleasure, Public Discourse: Reflections on Engaging with Literature.” In Language and Verbal Art Revisited. Ed. by Donna R. Miller and Monica Turci. London: Equinox. Honko 1995 Lauri Honko. “Multiformit ja pitkän eepoksen arvoitus.” Sananjalka, 37:117-45. Honko 1998 . Textualizing the Siri Epic. Folklore Fellows’ Communications, 264. Helsinki: Academia Scientiarum Fennica. Honko 2003 , in collaboration with Anneli Honko and Paul Hagu. The Maiden’s Death Song and The Great Wedding: Anne Vabarna’s Oral Twin Epic Written down by A. O. Väisnen. Folklore Fellows’ Communications, 281. Helsinki: Suomalainen Tiedeakatemia. PARALLELISM DYNAMICS I 479 Honko and Honko 1995 Lauri Honko and Anneli Honko. “Multiforms in Epic Composition.” In XIth Congress of the International Society for Folk-Narrative Research (ISFNR), January 6-12, 1995, Mysore, India. Papers. Mysore: Central Institute of Indian Languages. vol. II, pp. 207-40. Honko and Honko 1998 . “Multiforms in Epic Composition.” In The Epic. Oral and Written. Ed. by Lauri Honko, Jawaharlal Handoo, and John Miles Foley. Mysore: Central Institute of Indian Languages. pp. 31-79. Hopkins 1959 [1865] Gerard Manley Hopkins. “Poetic Diction.” In The Journals and Papers of Gerard Manley Hopkins. Ed. by H. House and G. Storey. London: Oxford University Press. pp. 84-85. Hymes 1981 Dell Hymes. “In Vain I Tried to Tell You:” Essays in Native American Ethnopoetics. Philadelphia: University of Pennsylvania Press. Institutio oratoria Fabius Quintilian, The Institutio oratoria of Quintilian I-IV. Ed. and trans. by H. E. Butler. London: William Heinemann, 1920-22. Jakobson 1981 Roman Jakobson. Selected Writings III: Poetry of Grammar and Grammar of Poetry. Ed. by Stephen Rudy. The Hague: Mouton. Jakobson 1981 [1960] . “Linguistics and Poetics.” In Selected Writings III: Poetry of Grammar and Grammar of Poetry. Ed. by Stephen Rudy. The Hague: Mouton. pp. 18-51. Jakobson 1981 [1966] . “Grammatical Parallelism and Its Russian Facet.” In Selected Writings III: Poetry of Grammar and Grammar of Poetry. Ed. by Stephen Rudy. The Hague: Mouton. pp. 98-135. Jakobson 1987 [1956] . “Two Aspects of Language and Two Types of Aphasic Disturbances”. In Roman Jakobson, Language in Literature. Ed. by Krystyna Pomorska and Stephen Rudy. Cambridge: Harvard University Press. pp. 95-114. Johnstone 1991 Barbara Johnstone. Repetition in Arabic Discourse: Paradigms, Syntagms and the Ecology of Language. Amsterdam: John Benjamins. Krohn 1918 Kaarle Krohn. Kalevalankysymyksia I-II. Suomalais-Ugrilaisen Seuran Aikakauskirja, 35-36. Helsinki: Suomalais-Urgilaisen Seura. Krohn 1926 . Die folkloristische Arbeitsmethode: Begründet von Julius Krohn und weitergeführt von nordischen Forschern. Oslo: Aschehoug. 480 FROG Kuusi 1949 Matti Kuusi. Sampo-eepos: Typologinen analyysi. Suomalais-Ugrilaisen Seuran Tutkimuksia, 96. Helsinki: Suomalais-Ugrilainen Seura. Kuusi 1994 . “Questions of Kalevala Meter: What Exactly Did Kalevala Language Signify to Its Users?” In Songs Beyond the Kalevala: Transformations of Oral Poetry. Ed. by Anna-Leena Siikala and Sinikka Vakimo. Studia Fennica Folkloristica, 2. Helsinki: Finnish Literature Society. pp. 41-55. Kuusi et al. 1977 Matti Kuusi, Keith Bosley, and Michael Branch, ed. and trans. Finnish Folk Poetry—Epic: An Anthology in Finnish and English. Helsinki: Finnish Literature Society. Lamb 2015 William Lamb. “Verbal Formulas in Gaelic Traditional Narrative: Some Aspects of Their Form and Function.” In Registers of Communication. Ed. by Asif Agha and Frog. Studia Fennica Linguistica, 18. Helsinki: Finnish Literature Society. pp. 225-46. Leino 1986 Pentti Leino. Language and Metre: Metrics and the Metrical System of Finnish. Studia Fennica, 31. Helsinki: Finnish Literature Society. Lord 1960 Albert B. Lord. The Singer of Tales. Cambridge: Harvard University Press. Lowth 1753 Robert Lowth. De sacra poesi Hebræorum: Prælectiones academic Oxonii habitæ a Roberto Lowth A. M. collegii novi nuper socio, et poeticæ public prælectore: Subjicitur metricae harianae brevis confutatio et oratio crewiana. Oxonius: Clarendonia. Lukin 2014 Karina Lukin. “Notes on Parallelism in Nenets Epic.” In Parallelism in Verbal Art and Performance: Pre-Print Papers of the Seminar-Workshop, 26th-27th May 2014. Folkloristiikan Toimite, 21. Helsinki: Folklore Studies, University of Helsinki. pp. 120-29. Mede 1653 Josephus Mede. Dissertationum ecclesiasticarum triga: De sanctitate relative, De veneratione sacra, De sortitione et alea: Quibus accednnt [sic] fragmenta sacra. Londini: Jo. Martin, Jac. Allestrye, et Tim. Garthwait. Morris 1971 [1938] Charles W. Morris. “Foundations of a Theory of Signs.” In Charles W. Morris, Writings on the General Theory of Signs. Ed. by Thomas A. Sebeok. Approaches to Semiotics, 16. The Hauge: Mouton. pp. 17-72. Oinas 1985 [1976] Felix J. Oinas, ed. “Negative Parallelism in Karelian-Finnish Folklore.” Studies in Finnic Folklore. Helsinki: Finnish Literature Society. pp. 78-92. PARALLELISM DYNAMICS I 481 Parry 1928 Milman Parry. L’épithète traditionnelle dans Homère. Paris: Société d’Éditions “Les Belles Lettres.” Porthan 1766 Henricus Gabriel Porthan. Disserttationis de poësi Fennica. Åbo: Joh. Christoph. Frenckell. Roper 2012 Jonathan Roper. “Synonymy and Rank in Alliterative Poetry.” Sign Systems Studies, 40:82-93. Saarinen 2013 Jukka Saarinen. “Behind the Text: Reconstructing the Voice of a Singer.” RMN Newsletter, 7:34-43. Saarinen 1994 . “The Päivölä Song of Miihkali Perttunen.” In Songs Beyond the Kalevala: Transformations of Oral Poetry. Studia Fennica Folkloristica, 2. Ed. by Anna-Leena Siikala and Sinikka Vakimo. Helsinki: Finnish Literature Society. pp. 180-96. Sadeniemi 1951 Matti Sadeniemi. Die Metrik des Kalevala-verses. Folklore Fellows’ Communications, 139. Helsinki: Academia Scientiarum Fennicae. Sands 2001 Tracy R. Sands. “‘Det kommo tvänne Dufvor...’: Doves, Ravens, and the Dead in Scandinavian Folk Tradition.” Scandinavian Studies, 73.3:349-74. Sarv 1999 Mari Sarv. “Regilaul: Clearing the Alliterative Haze”. Folklore: Electronic Journal of Folklore, 10:126-40. Saussure 1967 [1916] Ferdinand de Saussure. Cours de linguistique générale. Paris: Éditions Payot and Rivages. Setälä 1932 E. N. Setälä. Sammon arvoitus. Helsinki: Otava. Shore 2015 Susanna Shore. “Register in Systemic Functional Linguistics.” In Registers of Communication. Ed. by Asif Agha and Frog. Studia Fennica Linguistica, 18. Helsinki: Finnish Literature Society. pp. 54-76. Siikala 2002 Anna-Leena Siikala. Mythic Images and Shamanism: A Perspective on Kalevala Poetry. Folklore Fellows’ Communications, 280. Helsinki: Academia Scientiarum Fennica. Silverstein 1984 Michael Silverstein. “On the Pragmatic ‘Poetry’ of Prose.” In Meaning, Form and Use in Context. Ed. by Deborah Schiffrin. Washington D.C.: Georgetown University Press. pp. 181‒99. 482 FROG Silverstein 2003 . “Indexical Order and the Dialectics of Sociolinguistic Life.” Language & Communication, 23:193-229. Silverstein 2004 . “‘Cultural’ Concepts and the Nexus of Language.” Current Anthropology, 45:621-52. SKVR Suomen Kansan Vanhat Runot I-XV. Helsinki: Suomalaisen Kirjallisuuden Seura, 1908-97. Steinitz 1934 Wolfgang Steinitz. Der Parallelismus in der finnisch-karelischen Volksdichtung untersucht an den Liedern des karelischen Sangers Arhippa Perttunen. Folklore Fellows’ Communications, 115. Helsinki: Academia Scientiarum Fennica. Tarkka 2012 Lotte Tarkka. “The Sampo: Myth and Vernacular Imagination.” In Mythic Discourses: Studies in Uralic Traditions. Studia Fennica Folkloristica, 20. Ed. by Frog, Anna-Leena Siikala, and Eila Stepanova. Helsinki: Finnish Literature Society. pp. 143-70. Tarkka 2013 . Songs of the Border People: Genre, Reflexivity, and Performance in Karelian Oral Poetry. Folklore Fellows’ Communications, 305. Helsinki: Academia Scientiarum Fennica. Tedlock 1983 Dennis Tedlock. The Spoken Word and the Work of Interpretation. Philadelphia: University of Pennsylvania Press. Thompson 1955 Stith Thompson. Narrative Motif-Analysis as a Folklore Method. Folklore Fellows’ Communications, 161. Helsinki: Academia Scientiarum Fennica. Thompson 1955-58 . Motif-Index of Folk-Literature: A Classification of Narrative Elements in Folktales, Ballads, Myths, Fables, Mediaeval Romances, Exempla, Fabliaux, Jest-Books and Local Legends, I-VI. Revised and expanded edition. Copenhagen: Rosenkilde and Bagger. Tsur 1992 Reuven Tsur. Toward a Theory of Cognitive Poetics. Amsterdam: NorthHolland. Ullholm/Aurivillius 1758 Sven Ullholm (student)/Carol Aurivillius (professor). De poësi biblica. M.A. Thesis. Upsaliae: L. M. Höjer. Urban 1986 Greg Urban. “The Semiotic Functions of Macro-Parallelism in the Sholkeng Origin Myth.” In Native South American Discourse. Ed. by Joel Sherzer and Greg Urban. Berlin: Mouton de Gruyter. pp. 15-57. PARALLELISM DYNAMICS I 483 Urban 1991 . A Discourse-Centered Approach to Culture: Native South American Myths and Rituals. Austin: University of Texas Press. Wray 2008 Alison Wray. Formulaic Language: Pushing the Boundaries. Oxford: Oxford University Press. 484 FROG This page is intentionally left blank.